{
  "papers": [
    {
      "title": "Reset Method based on the Theory of Manifold Optimization on Real Manifolds",
      "authors": [
        "Weiping Liu",
        "Jiajun Wang",
        "He Li",
        "Youfa Liu",
        "Jingui Zou"
      ],
      "abstract": "Manifold optimization is prominent in the fields of applied mathematics, statistics, machine learning, and in particular, deep learning. By leveraging the intrinsic geometric properties of manifolds, constrained optimization problems can be transformed into unconstrained optimization problems on certain manifolds.  An innovative method, Reset Method, is introduced that combines manifold optimization and standard methods (SGD, Adam and AdamW), aiming to enhance the improvement of precision. The efficacy of our proposed method is corroborated by extensive deep learning experiments, providing visible higher precision.",
      "url": "https://openreview.net/forum?id=xVw8YNEtH3",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "xVw8YNEtH3",
      "track": "main",
      "status": "Reject",
      "keywords": "Manifold Optimization;Real Manifolds;Method;Deep Learning.",
      "tldr": "",
      "primary_area": "optimization",
      "similarity_score": 15.862464623765211,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 15.862464623765211,
      "combined_score": 0.0,
      "rank": 1
    },
    {
      "title": "Neural Implicit Manifold Learning for Topology-Aware Generative Modelling",
      "authors": [
        "Brendan Leigh Ross",
        "Gabriel Loaiza-Ganem",
        "Anthony L. Caterini",
        "Jesse C Cresswell"
      ],
      "abstract": "Natural data observed in $\\mathbb{R}^n$ is often constrained to an $m$-dimensional manifold $\\mathcal{M}$, where $m < n$. Current probabilistic models represent this manifold by mapping an $m$-dimensional latent variable through a neural network $f_\\theta: \\mathbb{R}^m \\to \\mathbb{R}^n$. Such procedures, which we call pushforward models, incur a straightforward limitation: manifolds cannot in general be represented with a single parameterization, meaning that attempts to do so will incur either computational instability or the inability to learn probability densities within the manifold. To remedy this problem, we propose to model $\\mathcal{M}$ as a neural implicit manifold: the set of zeros of a neural network. To learn the data distribution within $\\mathcal{M}$, we introduce constrained energy-based models, which use a constrained variant of Langevin dynamics to train and sample within a learned manifold. The resulting model can be manipulated with an arithmetic of manifolds, which allows practitioners to take unions and intersections of model manifolds. In experiments on synthetic and natural data, we show that constrained EBMs can learn manifold-supported distributions with complex topologies more accurately than pushforward models.",
      "url": "https://openreview.net/forum?id=WA35e2vPlFT",
      "year": 2023,
      "venue": "ICLR 2023",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "WA35e2vPlFT",
      "track": "main",
      "status": "Reject",
      "keywords": "Manifold Learning;Unsupervised Learning;Density Estimation;Topology;Differential Geometry;Generative Modelling",
      "tldr": "We propose a new model for probability distributions on topologically complex data manifolds which learns manifolds implicitly as the set of zeros of a neural network and then learns the distribution within using a constrained energy-based model.",
      "primary_area": "",
      "similarity_score": 15.827814730029807,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 15.827814730029807,
      "combined_score": 0.0,
      "rank": 2
    },
    {
      "title": "Learning a Manifold as an Atlas",
      "authors": [
        "Nikolaos Pitelis",
        "Chris Russell",
        "Lourdes Agapito"
      ],
      "abstract": "In this work, we return to the underlying mathematical definition of a manifold and directly characterise learning a manifold as finding an atlas, or a set of overlapping charts, that accurately describe local structure. We formulate the problem of learning the manifold as an optimisation that simultaneously refines the continuous parameters defining the charts, and the discrete assignment of points to charts. In contrast to existing methods, this direct formulation of a manifold does not require \"unwrapping\" the manifold into a lower dimensional space and allows us to learn closed manifolds of interest to vision, such as those corresponding to gait cycles or camera pose. We report state-ofthe-art results for manifold based nearest neighbour classification on vision datasets, and show how the same techniques can be applied to the 3D reconstruction of human motion from a single image.",
      "url": "https://openaccess.thecvf.com/content_cvpr_2013/html/Pitelis_Learning_a_Manifold_2013_CVPR_paper.html",
      "year": 2013,
      "venue": "CVPR 2013",
      "source": "offline_cvpr",
      "doi": null,
      "pdf_url": "https://openaccess.thecvf.com/content_cvpr_2013/papers/Pitelis_Learning_a_Manifold_2013_CVPR_paper.pdf",
      "citations": null,
      "categories": [],
      "id": "859646c55f",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 14.567100086419527,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 14.567100086419527,
      "combined_score": 0.0,
      "rank": 3
    },
    {
      "title": "Harpoon: Generalised Manifold Guidance for Conditional Tabular Diffusion",
      "authors": [],
      "abstract": "Generating tabular data under conditions is critical to applications requiring precise control over the generative process. Existing methods rely on training-time strategies that do not generalise to unseen constraints during inference, and struggle to handle conditional tasks beyond tabular imputation. While manifold theory offers a principled way to guide generation, current formulations are tied to specific inference-time objectives and are limited to continuous domains. We extend manifold theory to tabular data and expand its scope to handle diverse inference-time objectives. On this foundation, we introduce Harpoon, a  tabular diffusion method that guides unconstrained samples along the manifold geometry to satisfy diverse tabular conditions at inference. We validate our theoretical contributions empirically on tasks such as imputation and enforcing inequality constraints, demonstrating Harpoon's strong performance across diverse datasets and the practical benefits of manifold-aware guidance for tabular data. Code URL: https://anonymous.4open.science/r/ManifoldTabularImputation-44E4/",
      "url": "https://openreview.net/forum?id=G5g6tDg1ZE",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "G5g6tDg1ZE",
      "track": "main",
      "status": "Active",
      "keywords": "Diffusion models;Conditional generation;Tabular diffusion;Manifold learning",
      "tldr": "",
      "primary_area": "generative models",
      "similarity_score": 14.214760607748556,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 14.214760607748556,
      "combined_score": 0.0,
      "rank": 4
    },
    {
      "title": "Sampling in Constrained Domains with Orthogonal-Space Variational Gradient Descent",
      "authors": [
        "Ruqi Zhang",
        "qiang liu",
        "Xin T. Tong"
      ],
      "abstract": "Sampling methods, as important inference and learning techniques, are typically designed for unconstrained domains. However, constraints are ubiquitous in machine learning problems, such as those on safety, fairness, robustness, and many other properties that must be satisfied to apply sampling results in real-life applications. Enforcing these constraints often leads to implicitly-defined manifolds, making efficient sampling with constraints very challenging. In this paper, we propose a new variational framework with a designed orthogonal-space gradient flow (O-Gradient) for sampling on a manifold $\\mathcal{G}_0$ defined by general equality constraints. O-Gradient decomposes the gradient into two parts: one decreases the distance to $\\mathcal{G}_0$ and the other decreases the KL divergence in the orthogonal space. While most existing manifold sampling methods require initialization on $\\mathcal{G}_0$, O-Gradient does not require such prior knowledge. We prove that O-Gradient converges to the target constrained distribution with rate $\\widetilde{O}(1/\\text{the number of iterations})$ under mild conditions. Our proof relies on a new Stein characterization of conditional measure which could be of independent interest. We implement O-Gradient through both Langevin dynamics and Stein variational gradient descent and demonstrate its effectiveness in various experiments, including Bayesian deep neural networks.",
      "url": "https://nips.cc/virtual/2022/poster/53704",
      "year": 2022,
      "venue": "NIPS 2022",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=peFP9Pl-6-_",
      "citations": null,
      "categories": [],
      "id": "peFP9Pl-6-_",
      "track": "main",
      "status": "Accept",
      "keywords": "",
      "tldr": " We propose a variational framework for sampling in general constrained domains, with theoretical guarantees and practical algorithms.",
      "primary_area": "",
      "similarity_score": 14.071855966579731,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 14.071855966579731,
      "combined_score": 0.0,
      "rank": 5
    },
    {
      "title": "A Discussion On the Validity of Manifold Learning",
      "authors": [
        "Dai Shi",
        "Andi Han",
        "Yi Guo",
        "Junbin Gao"
      ],
      "abstract": "Dimensionality reduction (DR) and manifold learning (ManL) have been applied extensively in many machine learning tasks, including signal processing, speech recognition, and neuroinformatics. However, the understanding of whether DR and ManL models can generate valid learning results remains unclear. In this work, we investigate the validity of learning results of some widely used DR and ManL methods through the chart mapping function of a manifold. We identify a fundamental problem of these methods: the mapping functions induced by these methods violate the basic settings of manifolds, and hence they are not learning manifold in the mathematical sense. To address this problem, we provide a provably correct algorithm called fixed points Laplacian mapping (FPLM), that has the geometric guarantee to find a valid manifold representation (up to a homeomorphism). Combining one additional condition (orientation preserving), we discuss a sufficient condition for an algorithm to be bijective for any -simplex decomposition result on a -manifold.  However,  constructing such a mapping function and its computational method satisfying these conditions is still an open problem in mathematics.",
      "url": "https://openreview.net/forum?id=ad_F_z27pCx",
      "year": 2022,
      "venue": "ICLR 2022",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "ad_F_z27pCx",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Manifold learning;Dimensionality Reduction;Computational Geometry;Simplicial Complex",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 13.645393761208583,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.645393761208583,
      "combined_score": 0.0,
      "rank": 6
    },
    {
      "title": "Constrained Diffusion for Protein Design with Hard Structural Constraints",
      "authors": [],
      "abstract": "Diffusion models offer a powerful means of capturing the manifold of realistic protein structures, enabling rapid design for protein engineering tasks. However, existing approaches observe critical failure modes when precise constraints are necessary for functional design. To this end, we present a constrained diffusion framework for structure-guided protein design, ensuring strict adherence to functional requirements while maintaining precise stereochemical and geometric feasibility. The approach integrates proximal feasibility updates with ADMM decomposition into the generative process, scaling effectively to the complex constraint sets of this domain. We evaluate on challenging protein design tasks, including motif scaffolding and vacancy-constrained pocket design, while introducing a novel curated benchmark dataset for motif scaffolding in the PDZ domain. Our approach achieves state-of-the-art, providing perfect satisfaction of bonding and geometric constraints with no degradation in structural diversity.",
      "url": "https://openreview.net/forum?id=kkvqVRu2Zy",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "kkvqVRu2Zy",
      "track": "main",
      "status": "Active",
      "keywords": "Constrained Diffusion;Generative Models;Protein Design;Proximal Optimization;Motif Scaffolding",
      "tldr": "",
      "primary_area": "applications to physical sciences (physics, chemistry, biology, etc.)",
      "similarity_score": 13.645131509105356,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.645131509105356,
      "combined_score": 0.0,
      "rank": 7
    },
    {
      "title": "For Manifold Learning, Deep Neural Networks Can be Locality Sensitive Hash Functions",
      "authors": [
        "Nishanth Dikkala",
        "Gal Kaplun",
        "Rina Panigrahy"
      ],
      "abstract": "It is well established that training deep neural networks gives useful representations that capture essential features of the inputs. However, these representations are poorly understood in theory and practice. In the context of supervised learning an important question is whether these representations capture features informative for classification, while filtering out non-informative noisy ones. We present a formal framework to study this question by considering a generative process where each class is associated with a high-dimensional manifold and different classes define different manifolds. Under this model, each input is produced using two latent vectors: (i) a ``manifold identifier\"  $\\gamma$ and; (ii)~a ``transformation parameter\" $\\theta$ that shifts examples along the surface of a manifold. E.g., $\\gamma$ might represent a canonical image of a dog, and $\\theta$ might stand for variations in pose, background or lighting. We provide theoretical evidence that neural representations can be viewed as LSH-like functions that map each input to an embedding that is a function of solely the informative $\\gamma$ and invariant to $\\theta$, effectively recovering the manifold identifier . We formally show that we get one-shot learning to unseen classes as an important consequence of this behavior.",
      "url": "https://openreview.net/forum?id=ZTZa78mCbie",
      "year": 2022,
      "venue": "ICLR 2022",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "ZTZa78mCbie",
      "track": "main",
      "status": "Withdraw",
      "keywords": "theory of deep learning;theory of representation learning;manifold learning;locality sensitive hash functions;interpretability",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 13.609456935297061,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.609456935297061,
      "combined_score": 0.0,
      "rank": 8
    },
    {
      "title": "An Information Geometry of Statistical Manifold Learning",
      "authors": [
        "Ke Sun",
        "Stéphane Marchand-Maillet"
      ],
      "abstract": "Manifold learning seeks low-dimensional representations of high-dimensional data. The main tactics have been exploring the geometry in an input data space and an output embedding space. We develop a manifold learning theory in a hypothesis space consisting of models. A model means a specific instance of a collection of points, e.g., the input data collectively or the output embedding collectively. The semi-Riemannian metric of this hypothesis space is uniquely derived in closed form based on the information geometry of probability distributions. There, manifold learning is interpreted as a trajectory of intermediate models. The volume of a continuous region reveals an amount of information. It can be measured to define model complexity and embedding quality. This provides deep unified perspectives of manifold learning theory.",
      "url": "https://proceedings.mlr.press/v32/suna14.html",
      "year": 2014,
      "venue": "ICML 2014",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "http://proceedings.mlr.press/v32/suna14.pdf",
      "citations": null,
      "categories": [],
      "id": "40f1211589",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 13.588387224146109,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.588387224146109,
      "combined_score": 0.0,
      "rank": 9
    },
    {
      "title": "Constrained Diffusion with Trust Sampling",
      "authors": [
        "William Huang",
        "Yifeng Jiang",
        "Tom Van Wouwe",
        "Karen Liu"
      ],
      "abstract": "Diffusion models have demonstrated significant promise in various generative tasks; however, they often struggle to satisfy challenging constraints. Our approach addresses this limitation by rethinking training-free loss-guided diffusion from an optimization perspective. We formulate a series of constrained optimizations throughout the inference process of a diffusion model. In each optimization, we allow the sample to take multiple steps along the gradient of the proxy constraint function until we can no longer trust the proxy, according to the variance at each diffusion level. Additionally, we estimate the state manifold of diffusion model to allow for early termination when the sample starts to wander away from the state manifold at each diffusion step. Trust sampling effectively balances between following the unconditional diffusion model and adhering to the loss guidance, enabling more flexible and accurate constrained generation.  We demonstrate the efficacy of our method through extensive experiments on complex tasks, and in drastically different domains of images and 3D motion generation, showing significant improvements over existing methods in terms of generation quality.  Our implementation is available at https://github.com/will-s-h/trust-sampling.",
      "url": "https://neurips.cc/virtual/2024/poster/94344",
      "year": 2024,
      "venue": "NIPS 2024",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=dJUb9XRoZI",
      "citations": null,
      "categories": [],
      "id": "dJUb9XRoZI",
      "track": "main",
      "status": "Poster",
      "keywords": "diffusion models;guidance;image generation;human motion",
      "tldr": "",
      "primary_area": "diffusion_based_models",
      "similarity_score": 13.571114683906725,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.571114683906725,
      "combined_score": 0.0,
      "rank": 10
    },
    {
      "title": "Validating the Lottery Ticket Hypothesis with Inertial Manifold Theory",
      "authors": [
        "Zeru Zhang",
        "Jiayin Jin",
        "Zijie Zhang",
        "Yang Zhou",
        "Xin Zhao",
        "Jiaxiang Ren",
        "Ji Liu",
        "Lingfei Wu",
        "Ruoming Jin",
        "Dejing Dou"
      ],
      "abstract": "Despite achieving remarkable efficiency, traditional network pruning techniques often follow manually-crafted heuristics to generate pruned sparse networks. Such heuristic pruning strategies are hard to guarantee that the pruned networks achieve test accuracy comparable to the original dense ones. Recent works have empirically identified and verified the Lottery Ticket Hypothesis (LTH): a randomly-initialized dense neural network contains an extremely sparse subnetwork, which can be trained to achieve similar accuracy to the former. Due to the lack of theoretical evidence, they often need to run multiple rounds of expensive training and pruning over the original large networks to discover the sparse subnetworks with low accuracy loss. By leveraging dynamical systems theory and inertial manifold theory, this work theoretically verifies the validity of the LTH. We explore the possibility of theoretically lossless pruning as well as one-time pruning, compared with existing neural network pruning and LTH techniques. We reformulate the neural network optimization problem as a gradient dynamical system and reduce this high-dimensional system onto inertial manifolds to obtain a low-dimensional system regarding pruned subnetworks. We demonstrate the precondition and existence of pruned subnetworks and prune the original networks in terms of the gap in their spectrum that make the subnetworks have the smallest dimensions.",
      "url": "https://nips.cc/virtual/2021/poster/26292",
      "year": 2021,
      "venue": "NIPS 2021",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=h6EWbx5xTj7",
      "citations": null,
      "categories": [],
      "id": "h6EWbx5xTj7",
      "track": "main",
      "status": "Poster",
      "keywords": "Lottery Ticket Hypothesis;neural network pruning;dynamical systems;inertial manifold;theoretical evidence",
      "tldr": "Theoretically verify the precondition and validity of the Lottery Ticket Hypothesis",
      "primary_area": "",
      "similarity_score": 13.527295228071175,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.527295228071175,
      "combined_score": 0.0,
      "rank": 11
    },
    {
      "title": "Enhancing Cross-Lingual and Cross-Domain Adaptability in Large Language Models for Software Engineering",
      "authors": [
        "Yuanhao Li",
        "Haocheng Yang",
        "Wei Tan",
        "Shilong Yuan",
        "Hongbo Wang",
        "Zhenghan chen"
      ],
      "abstract": "This paper presents a groundbreaking mathematical framework for unsupervised domain adaptation (UDA) in the context of cross-lingual and cross-domain code modeling. We introduce the Enhanced Dynamic Code Modeling (UDA-EDCM) system, which leverages advanced concepts from measure theory, differential geometry, and information geometry to address the challenges posed by the diversity of natural and programming languages. At the core of UDA-EDCM is a novel measure-theoretic formulation of domain adaptation, utilizing optimal transport theory to minimize the discrepancy between source and target domains. We develop a Riemannian manifold approach to feature space alignment, introducing a Geodesic Flow Kernel that captures the intrinsic geometry of the code representation space. The UDA-EDCM operator is analyzed through the lens of functional analysis, revealing its spectral properties and their implications for generalization. Our information-theoretic bound on domain adaptation provides insights into the fundamental limits of knowledge transfer in code modeling. We present a unified theorem that synthesizes these diverse mathematical perspectives, offering a comprehensive characterization of UDA-EDCM's performance in terms of Wasserstein distance, empirical Rademacher complexity, and Fisher information. This theoretical foundation is complemented by an innovative optimization framework based on the Fisher Information Metric, ensuring efficient convergence in the probabilistic manifold of model parameters. Extensive experiments demonstrate that UDA-EDCM significantly outperforms existing approaches in zero-shot and few-shot learning scenarios across a wide range of programming languages and coding tasks. Our work not only advances the baselines in domain adaptation for code intelligence but also establishes a rigorous mathematical basis for future research in adaptive AI systems for software engineering.",
      "url": "https://openreview.net/forum?id=XFCKEgGhEK",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "XFCKEgGhEK",
      "track": "main",
      "status": "Reject",
      "keywords": "Code Generation;Transfer learning",
      "tldr": "",
      "primary_area": "transfer learning, meta learning, and lifelong learning",
      "similarity_score": 13.238157697070676,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.238157697070676,
      "combined_score": 0.0,
      "rank": 12
    },
    {
      "title": "RoSE: Enhancing SE(3)-based Protein Backbone Generation via Robust Score Estimation",
      "authors": [
        "Minzhang Li",
        "Haochen Wang",
        "Weichen Qin",
        "Yifan Qin",
        "Jiakai Zhang",
        "Jiayi Dou",
        "Jingyi Yu"
      ],
      "abstract": "This work presents improvements to Riemannian diffusion models for protein structure generation by developing robust heat kernel computation methods on $SE(3)$ space. While existing approaches suffer from approximation errors in score-based diffusion, our method enables stable and accurate denoising score matching on the high-dimensional $SE(3)^N$ manifold through theoretically-grounded numerical techniques. The proposed framework achieves competitive performance in protein generation benchmarks, demonstrating superior scores and successfully generating diverse, physically-plausible protein structures. Notably, our model solves 23 out of 24 motif scaffolding problems and designs refoldable nanobodies, significantly advancing the capability to generate functional protein geometries while maintaining mathematical consistency with the underlying manifold structure.",
      "url": "https://openreview.net/forum?id=yV2bsMVfal",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "yV2bsMVfal",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Riemannian Diffusion;Protein Design",
      "tldr": "",
      "primary_area": "applications to physical sciences (physics, chemistry, biology, etc.)",
      "similarity_score": 13.208870708515017,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.208870708515017,
      "combined_score": 0.0,
      "rank": 13
    },
    {
      "title": "More About VLAD: A Leap From Euclidean to Riemannian Manifolds",
      "authors": [
        "Masoud Faraki",
        "Mehrtash T. Harandi",
        "Fatih Porikli"
      ],
      "abstract": "This paper takes a step forward in image and video coding by extending the well-known Vector of Locally Aggregated Descriptors (VLAD) onto an extensive space of curved Riemannian manifolds. We provide a comprehensive mathematical framework that formulates the aggregation problem of such manifold data into an elegant solution. In particular, we consider structured descriptors from visual data, namely Region Covariance Descriptors and linear subspaces that reside on the manifold of Symmetric Positive Definite matrices and the Grassmannian manifolds, respectively. Through rigorous experimental validation, we demonstrate the superior performance of this novel Riemannian VLAD descriptor on several visual classification tasks including video-based face recognition, dynamic scene recognition, and head pose classification.",
      "url": "https://openaccess.thecvf.com/content_cvpr_2015/html/Faraki_More_About_VLAD_2015_CVPR_paper.html",
      "year": 2015,
      "venue": "CVPR 2015",
      "source": "offline_cvpr",
      "doi": null,
      "pdf_url": "https://openaccess.thecvf.com/content_cvpr_2015/papers/Faraki_More_About_VLAD_2015_CVPR_paper.pdf",
      "citations": null,
      "categories": [],
      "id": "1222a1b772",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 13.180931372481597,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.180931372481597,
      "combined_score": 0.0,
      "rank": 14
    },
    {
      "title": "Manifold structure in graph embeddings",
      "authors": [
        "Patrick Rubin-Delanchy"
      ],
      "abstract": "Statistical analysis of a graph often starts with embedding, the process of representing its nodes as points in space. How to choose the embedding dimension is a nuanced decision in practice, but in theory a notion of true dimension is often available. In spectral embedding, this dimension may be very high. However, this paper shows that existing random graph models, including graphon and other latent position models, predict the data should live near a much lower-dimensional set. One may therefore circumvent the curse of dimensionality by employing methods which exploit hidden manifold structure.",
      "url": "https://nips.cc/virtual/2020/poster/17864",
      "year": 2020,
      "venue": "NIPS 2020",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2020/file/8682cc30db9c025ecd3fee433f8ab54c-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "17864",
      "track": "main",
      "status": "Spotlight",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 13.164215769583805,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.164215769583805,
      "combined_score": 0.0,
      "rank": 15
    },
    {
      "title": "A Manifold Perspective on the Statistical Generalization of Graph Neural Networks",
      "authors": [
        "Zhiyang Wang",
        "Juan Cervino",
        "Alejandro Ribeiro"
      ],
      "abstract": "Graph Neural Networks (GNNs) extend convolutional neural networks to operate on graphs. Despite their impressive performances in various graph learning tasks, the theoretical understanding of their generalization capability is still lacking. Previous GNN generalization bounds ignore the underlying graph structures, often leading to bounds that increase with the number of nodes – a behavior contrary to the one experienced in practice. In this paper, we take a manifold perspective to establish the statistical generalization theory of GNNs on graphs sampled from a manifold in the spectral domain. As demonstrated empirically, we prove that the generalization bounds of GNNs decrease linearly with the size of the graphs in the logarithmic scale, and increase linearly with the spectral continuity constants of the filter functions. Notably, our theory explains both node-level and graph-level tasks. Our result has two implications: i) guaranteeing the generalization of GNNs to unseen data over manifolds; ii) providing insights into the practical design of GNNs, i.e., restrictions on the discriminability of GNNs are necessary to obtain a better generalization performance. We demonstrate our generalization bounds of GNNs using synthetic and multiple real-world datasets.",
      "url": "https://icml.cc/virtual/2025/poster/46337",
      "year": 2025,
      "venue": "ICML 2025",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=7Cxk63lTTm",
      "citations": null,
      "categories": [],
      "id": "7Cxk63lTTm",
      "track": "main",
      "status": "Poster",
      "keywords": "graph neural networks;generalization",
      "tldr": "",
      "primary_area": "deep_learning->graph_neural_networks",
      "similarity_score": 13.115024327488882,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.115024327488882,
      "combined_score": 0.0,
      "rank": 16
    },
    {
      "title": "S$^2$MAM: Semi-supervised Meta Additive Model for Robust Estimation and Variable Selection",
      "authors": [
        "Xuelin Zhang",
        "Hong Chen",
        "Yingjie Wang",
        "Zeyu Zhang",
        "Tieliang Gong",
        "Bin Gu",
        "Feng Zheng"
      ],
      "abstract": "Semi-supervised learning with manifold regularization is a classical family for learning from the labeled and unlabeled data jointly, where the key requirement is the support of unknown marginal distribution enjoys the geometric structure of a Riemannian manifold. Usually, the Laplace-Beltrami operator-based manifold regularization can be approximated empirically by the Laplacian regularization associated with the whole training data and its graph Laplacian matrix. However, the graph Laplacian matrix depends heavily on the pre-specifying similarity metric and may result in inappropriate penalties when facing redundant and noisy input variables. In order to address the above issues, this paper proposes a new semi-supervised meta additive model (S$^2$MAM) under a bilevel optimization scheme to automatically identify the informative variables, update the similarity matrix, and achieve the interpretable prediction simultaneously. Theoretical guarantees are provided for S$^2$MAM including the computing convergence and the statistical generalization bound. Experimental assessments on synthetic and real-world datasets validate the robustness and interpretability of the proposed approach.",
      "url": "https://openreview.net/forum?id=dpnPOXoqVQ",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "dpnPOXoqVQ",
      "track": "main",
      "status": "Reject",
      "keywords": "manifold regularization;bilevel optimization;sparse additive model;robustness;learning theory",
      "tldr": "",
      "primary_area": "learning theory",
      "similarity_score": 13.010128597456504,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 13.010128597456504,
      "combined_score": 0.0,
      "rank": 17
    },
    {
      "title": "Constrained Reinforcement Learning using Bender’s Decomposition and Exact Constraint Satisfaction",
      "authors": [
        "Alexander Mattick",
        "Christopher Mutschler"
      ],
      "abstract": "Recent advancements in reinforcement learning (RL) have expanded its applications beyond sequential decision-making to encompass non-sequential tasks, such as matrix decompositions, automatic generation of sorting networks, and combinatorial optimization. However, these tasks often require problem-specific algorithm designs to ensure the validity of the solution.\nTo address this limitation, we propose a universal framework that reformulates non-sequential tasks as constrained RL problems by learning to generate cutting planes, i.e., mathematical constraints that systematically refine the solution space. We ensure constraint satisfaction throughout the training process, enabling safe and efficient training even during deployment.\nWe show the efficacy of our framework on two complex optimization problems: a reward-maximizing stochastic job-shop scheduling problem and a nonlinear, nonconvex packing problem. Our method achieves near-globally optimal solutions while accelerating convergence by up to a factor of 800.",
      "url": "https://openreview.net/forum?id=KJ3zkHzsKm",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "KJ3zkHzsKm",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Reinforcement Learning;Constrained Reinforcement Learning;Optimization",
      "tldr": "",
      "primary_area": "reinforcement learning",
      "similarity_score": 12.998139045639963,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.998139045639963,
      "combined_score": 0.0,
      "rank": 18
    },
    {
      "title": "A Statistical Manifold Framework for Point Cloud Data",
      "authors": [
        "Yonghyeon Lee",
        "Seungyeon Kim",
        "Jinwon Choi",
        "Frank C. Park"
      ],
      "abstract": "A large class of problems in machine learning involve data sets in which each data point is a point cloud in $\\mathbb{R}^D$. The reason that most machine learning algorithms designed for point cloud data tend to be ad hoc, and difficult to measure their performance in a uniform and quantitative way, can be traced to the lack of a rigorous mathematical characterization of this space of point cloud data. The primary contribution of this paper is a Riemannian geometric structure for point cloud data. By interpreting the point cloud data as a set of samples from some underlying probability distribution, the set of point cloud data can be given the structure of a statistical manifold, with the Fisher information metric acting as a natural Riemannian metric; this structure then leads to, e.g., distance metrics, volume forms, and other coordinate-invariant, geometrically well-defined measures needed for applications. The only requirement on the part of the user is the choice of a meaningful underlying probability distribution, which is more intuitive and natural to make than what is required in existing ad hoc formulations.  Two autoencoder case studies involving point cloud data are presented to demonstrate the advantages of our statistical manifold framework: (i) interpolating between two 3D point cloud data sets to smoothly deform one object into another; (ii) transforming the latent coordinates into another with less distortion. Experiments with synthetic and large-scale standard benchmark point cloud data show more natural and intuitive shape evolutions, and improved classification accuracy for linear SVM vis-\\`{a}-vis existing methods.",
      "url": "https://openreview.net/forum?id=Tubzedlc4P",
      "year": 2022,
      "venue": "ICLR 2022",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "Tubzedlc4P",
      "track": "main",
      "status": "Reject",
      "keywords": "Riemannian Geometry;Point Cloud;Autoencoders",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 12.921407632633118,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.921407632633118,
      "combined_score": 0.0,
      "rank": 19
    },
    {
      "title": "Hyperbolic Binary Neural Network",
      "authors": [
        "Jun Chen",
        "Jingyang Xiang",
        "Tianxin Huang",
        "Xiangrui Zhao",
        "Yong Liu"
      ],
      "abstract": "Binary Neural Network (BNN) converts the full-precision weights and activations to the extreme 1-bit counterparts, which is especially suitable to be deployed on lightweight mobile devices. Neural network binarization is usually formulated as a constrained optimization problem, which restricts its optimized potential. In this paper, we introduce the dynamic exponential map that converts a constrained problem in the Riemannian manifold into an unconstrained one in the Euclidean space. Specifically, we propose a Hyperbolic Binary Neural Network (HBNN) by representing the parameter vector in the Euclidean space as the one in the hyperbolic space, which would enable us to optimize the parameter in an unconstrained space. By analyzing the parameterized representation, we present that the dynamic exponential map is a diffeomorphism in the Poincaré ball. Theoretically, this property will not create extra saddle points or local minima in the Poincaré ball, which also explains the good performance of the HBNN. Experiments on CIFAR10, CIFAR100, and ImageNet classification datasets with VGGsmall, ResNet18, and ResNet34 demonstrate the superiorities of our HBNN over existing state-of-the-art methods.",
      "url": "https://openreview.net/forum?id=Lv3MfAEgvVv",
      "year": 2023,
      "venue": "ICLR 2023",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "Lv3MfAEgvVv",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Neural network quantization;Hyperbolic geometry;Riemannian manifold",
      "tldr": "We propose a Hyperbolic Binary Neural Network that updates the parameters in hyperbolic space.",
      "primary_area": "",
      "similarity_score": 12.912407078823088,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.912407078823088,
      "combined_score": 0.0,
      "rank": 20
    },
    {
      "title": "Chart Auto-Encoders for Manifold Structured Data",
      "authors": [
        "Stephan Schonsheck",
        "Jie Chen",
        "Rongjie Lai"
      ],
      "abstract": " Auto-encoding and generative models have made tremendous successes in image and signal representation learning and generation. These models, however, generally employ the full Euclidean space or a bounded subset (such as $[0,1]^l$) as the latent space, whose trivial geometry is often too simplistic to meaningfully reflect the structure of the data. This paper aims at exploring a nontrivial geometric structure of the latent space for better data representation. Inspired by differential geometry, we propose \\textbf{Chart Auto-Encoder (CAE)}, which captures the manifold structure of the data with multiple charts and transition functions among them. CAE translates the mathematical definition of manifold through parameterizing the entire data set as a collection of overlapping charts, creating local latent representations. These representations are an enhancement of the single-charted latent space commonly employed in auto-encoding models, as they reflect the intrinsic structure of the manifold.  Therefore, CAE achieves a more accurate approximation of data and generates realistic new ones. We conduct experiments with synthetic and real-life data to demonstrate the effectiveness of the proposed CAE. ",
      "url": "https://openreview.net/forum?id=rJeBJJBYDB",
      "year": 2020,
      "venue": "ICLR 2020",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "rJeBJJBYDB",
      "track": "main",
      "status": "Reject",
      "keywords": "Auto-encoder;differential manifolds;multi-charted latent space",
      "tldr": "Manifold-structured latent space for generative models",
      "primary_area": "",
      "similarity_score": 12.894558206998033,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.894558206998033,
      "combined_score": 0.0,
      "rank": 21
    },
    {
      "title": "Positive Curvature and Hamiltonian Monte Carlo",
      "authors": [
        "Christof Seiler",
        "Simon Rubinstein-Salzedo",
        "Susan Holmes"
      ],
      "abstract": "The Jacobi metric introduced in mathematical physics can be used to analyze Hamiltonian Monte Carlo (HMC). In a geometrical setting, each step of HMC corresponds to a geodesic on a Riemannian manifold with a Jacobi metric. Our calculation of the sectional curvature of this HMC manifold allows us to see that it is positive in cases such as sampling from a high dimensional multivariate Gaussian. We show that positive curvature can be used to prove theoretical concentration results for HMC Markov chains.",
      "url": "https://nips.cc/virtual/2014/poster/4723",
      "year": 2014,
      "venue": "NIPS 2014",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2014/file/c76d3b26eba4f2c2fed5695faaae774f-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "4723",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 12.886607053331751,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.886607053331751,
      "combined_score": 0.0,
      "rank": 22
    },
    {
      "title": "Rethinking Large Language Model Distillation: A Constrained Markov Decision Process Perspective",
      "authors": [
        "Matthieu Zimmer",
        "Xiaotong Ji",
        "Tu Nguyen",
        "Haitham Bou Ammar"
      ],
      "abstract": "We introduce a novel approach to large language model (LLM) distillation by formulating it as a constrained reinforcement learning problem. While recent work has begun exploring the integration of task-specific rewards into distillation processes, existing methods typically rely on ad-hoc reward weighting.\nWe propose a principled optimization framework that maximizes task-specific rewards while constraining the divergence from the teacher model to remain below a specified threshold. Our approach adapts constrained state augmented reinforcement learning to the distillation setting, introducing a modified reward function that maintains theoretical guarantees of constraint satisfaction without requiring state augmentation or teacher model access during deployment and without the computational overhead of the dual Lagrangian methods. Through extensive experiments on mathematical reasoning tasks, we demonstrate that our method achieves better constraint satisfaction rates and better reasoning compared to the soft Lagrangian relaxation baselines while maintaining competitive task performance. Our framework provides a theoretically grounded and practically efficient solution for reward-aware distillation in resource-constrained settings.",
      "url": "https://openreview.net/forum?id=TBJIf2M23q",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "TBJIf2M23q",
      "track": "main",
      "status": "Withdraw",
      "keywords": "distillation;reinforcement learning;large language models;constrained reinforcement learning",
      "tldr": "",
      "primary_area": "foundation or frontier models, including LLMs",
      "similarity_score": 12.746404995119747,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.746404995119747,
      "combined_score": 0.0,
      "rank": 23
    },
    {
      "title": "Manifold Mixup: Better Representations by Interpolating Hidden States",
      "authors": [
        "Vikas Verma",
        "Alex Lamb",
        "Christopher Beckham",
        "Amir Najafi",
        "Ioannis Mitliagkas",
        "David Lopez-Paz",
        "Yoshua Bengio"
      ],
      "abstract": "Deep neural networks excel at learning the training data, but often provide incorrect and confident predictions when evaluated on slightly different test examples. This includes distribution shifts, outliers, and adversarial examples. To address these issues, we propose \\manifoldmixup{}, a simple regularizer that encourages neural networks to predict less confidently on interpolations of hidden representations. \\manifoldmixup{} leverages semantic interpolations as additional training signal, obtaining neural networks with smoother decision boundaries at multiple levels of representation. As a result, neural networks trained with \\manifoldmixup{} learn flatter class-representations, that is, with fewer directions of variance. We prove theory on why this flattening happens under ideal conditions, validate it empirically on practical situations, and connect it to the previous works on information theory and generalization. In spite of incurring no significant computation and being implemented in a few lines of code, \\manifoldmixup{} improves strong baselines in supervised learning, robustness to single-step adversarial attacks, and test log-likelihood.",
      "url": "https://icml.cc/virtual/2019/poster/3776",
      "year": 2019,
      "venue": "ICML 2019",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "http://proceedings.mlr.press/v97/verma19a/verma19a.pdf",
      "citations": null,
      "categories": [],
      "id": "3776",
      "track": "main",
      "status": "Oral",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 12.725678729478965,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.725678729478965,
      "combined_score": 0.0,
      "rank": 24
    },
    {
      "title": "Sample complexity and effective dimension for regression on manifolds",
      "authors": [
        "Andrew McRae",
        "Justin Romberg",
        "Mark Davenport"
      ],
      "abstract": "We consider the theory of regression on a manifold using reproducing kernel Hilbert space methods. Manifold models arise in a wide variety of modern machine learning problems, and our goal is to help understand the effectiveness of various implicit and explicit dimensionality-reduction methods that exploit manifold structure. Our first key contribution is to establish a novel nonasymptotic version of the Weyl law from differential geometry. From this we are able to show that certain spaces of smooth functions on a manifold are effectively finite-dimensional, with a complexity that scales according to the manifold dimension rather than any ambient data dimension. Finally, we show that given (potentially noisy) function values taken uniformly at random over a manifold, a kernel regression estimator (derived from the spectral decomposition of the manifold) yields minimax-optimal error bounds that are controlled by the effective dimension.",
      "url": "https://nips.cc/virtual/2020/poster/18845",
      "year": 2020,
      "venue": "NIPS 2020",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2020/file/977f8b33d303564416bf9f4ab1c39720-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "18845",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 12.591812722705232,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.591812722705232,
      "combined_score": 0.0,
      "rank": 25
    },
    {
      "title": "Constrained Binary Decision Making",
      "authors": [
        "Daniel Průša",
        "Vojtech Franc"
      ],
      "abstract": "Binary statistical decision making involves choosing between two states based on statistical evidence. The optimal decision strategy is typically formulated through a constrained optimization problem, where both the objective and constraints are expressed as integrals involving two Lebesgue measurable functions, one of which represents the strategy being optimized. In this work, we present a comprehensive formulation of the binary decision making problem and provide a detailed characterization of the optimal solution. Our framework encompasses a wide range of well-known and recently proposed decision making problems as specific cases. We demonstrate how our generic approach can be used to derive the optimal decision strategies for these diverse instances. Our results offer a robust mathematical tool that simplifies the process of solving both existing and novel formulations of binary decision making problems which are in the core of many Machine Learning algorithms.",
      "url": "https://neurips.cc/virtual/2024/poster/93660",
      "year": 2024,
      "venue": "NIPS 2024",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=ntV5xZfzEk",
      "citations": null,
      "categories": [],
      "id": "ntV5xZfzEk",
      "track": "main",
      "status": "Poster",
      "keywords": "binary statistical decision making;constrained optimization;Neyman-Pearson problem;selective classification",
      "tldr": "",
      "primary_area": "probabilistic_methods",
      "similarity_score": 12.580847918841936,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.580847918841936,
      "combined_score": 0.0,
      "rank": 26
    },
    {
      "title": "Learning Graph Convolution Filters from Data Manifold",
      "authors": [
        "Guokun Lai",
        "Hanxiao Liu",
        "Yiming Yang"
      ],
      "abstract": "Convolution Neural Network (CNN) has gained tremendous success in computer vision tasks with its outstanding ability to capture the local latent features. Recently, there has been an increasing interest in extending CNNs to the general spatial domain. Although various types of graph convolution and geometric convolution methods have been proposed, their connections to traditional 2D-convolution are not well-understood. In this paper, we show that depthwise separable convolution is a path to unify the two kinds of convolution methods in one mathematical view, based on which we derive a novel Depthwise Separable Graph Convolution that subsumes existing graph convolution methods as special cases of our formulation. Experiments show that the proposed approach consistently outperforms other graph convolution and geometric convolution baselines on benchmark datasets in multiple domains.",
      "url": "https://openreview.net/forum?id=H139Q_gAW",
      "year": 2018,
      "venue": "ICLR 2018",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "H139Q_gAW",
      "track": "main",
      "status": "Reject",
      "keywords": "Label Propagation;Depthwise separable convolution;Graph and geometric convolution",
      "tldr": "We devise a novel Depthwise Separable Graph Convolution (DSGC) for the generic spatial domain data, which is highly compatible with depthwise separable convolution.",
      "primary_area": "",
      "similarity_score": 12.52324097582431,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.52324097582431,
      "combined_score": 0.0,
      "rank": 27
    },
    {
      "title": "DeepMAD: Mathematical Architecture Design for Deep Convolutional Neural Network",
      "authors": [
        "Xuan Shen",
        "Yaohua Wang",
        "Ming Lin",
        "Yilun Huang",
        "Hao Tang",
        "Xiuyu Sun",
        "Yanzhi Wang"
      ],
      "abstract": "The rapid advances in Vision Transformer (ViT) refresh the state-of-the-art performances in various vision tasks, overshadowing the conventional CNN-based models. This ignites a few recent striking-back research in the CNN world showing that pure CNN models can achieve as good performance as ViT models when carefully tuned. While encouraging, designing such high-performance CNN models is challenging, requiring non-trivial prior knowledge of network design. To this end, a novel framework termed Mathematical Architecture Design for Deep CNN (DeepMAD) is proposed to design high-performance CNN models in a principled way. In DeepMAD, a CNN network is modeled as an information processing system whose expressiveness and effectiveness can be analytically formulated by their structural parameters. Then a constrained mathematical programming (MP) problem is proposed to optimize these structural parameters. The MP problem can be easily solved by off-the-shelf MP solvers on CPUs with a small memory footprint. In addition, DeepMAD is a pure mathematical framework: no GPU or training data is required during network design. The superiority of DeepMAD is validated on multiple large-scale computer vision benchmark datasets. Notably on ImageNet-1k, only using conventional convolutional layers, DeepMAD achieves 0.7% and 1.5% higher top-1 accuracy than ConvNeXt and Swin on Tiny level, and 0.8% and 0.9% higher on Small level.",
      "url": "https://cvpr.thecvf.com/virtual/2023/poster/21803",
      "year": 2023,
      "venue": "CVPR 2023",
      "source": "offline_cvpr",
      "doi": null,
      "pdf_url": "https://openaccess.thecvf.com/content/CVPR2023/papers/Shen_DeepMAD_Mathematical_Architecture_Design_for_Deep_Convolutional_Neural_Network_CVPR_2023_paper.pdf",
      "citations": null,
      "categories": [],
      "id": "21803",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 12.39613668148204,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.39613668148204,
      "combined_score": 0.0,
      "rank": 28
    },
    {
      "title": "Curvature Enhanced Manifold Sampling",
      "authors": [
        "Ilya Kaufman",
        "Omri Azencot"
      ],
      "abstract": "Over-parameterized deep learning models, characterized by their large number of parameters, have demonstrated remarkable performance in various tasks. Despite the potential risk of overfitting, these models often generalize well to unseen data due to effective regularization techniques, with data augmentation being one of the most prominent methods. This strategy has proven effective in classification tasks, where label-preserving transformations are applicable. However, the application of data augmentation in regression problems remains underexplored. Recently, a new *manifold learning* approach for sampling synthetic data has been introduced, and it can be viewed as utilizing a first-order approximation of the data manifold. In this work, we propose to extend this direction by providing the fundamental theory and practical tools for approximating and sampling general data manifolds. Further, we introduce the curvature enhanced manifold sampling (CEMS) data augmentation method for regression. CEMS is based on a second-order encoding of the manifold, facilitating sampling and reconstruction of new points. Through extensive evaluations on multiple datasets and in comparison to several state-of-the-art approaches, we demonstrate that CEMS is superior in in-distribution and out-of-distribution tasks, while incurring only a mild computational overhead.",
      "url": "https://openreview.net/forum?id=HYWdlCPtao",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "HYWdlCPtao",
      "track": "main",
      "status": "Reject",
      "keywords": "Manifold learning;Data augmentation;Regression",
      "tldr": "",
      "primary_area": "other topics in machine learning (i.e., none of the above)",
      "similarity_score": 12.37242649046323,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.37242649046323,
      "combined_score": 0.0,
      "rank": 29
    },
    {
      "title": "Encoded Prior Sliced Wasserstein AutoEncoder for learning latent manifold representations",
      "authors": [
        "Sanjukta Krishnagopal",
        "Jacob Bedrossian"
      ],
      "abstract": "While variational autoencoders have been successful in a variety of tasks, the use of conventional Gaussian or Gaussian mixture priors are limited in their ability to encode underlying structure of data in the latent representation.\nIn this work, we introduce an Encoded Prior Sliced Wasserstein AutoEncoder (EPSWAE) wherein an additional prior-encoder network facilitates learns an embedding of the data manifold which preserves topological and geometric properties of the data, thus improving the structure of latent space.\nThe autoencoder and prior-encoder networks are iteratively trained using the Sliced Wasserstein (SW) distance, which efficiently measures the distance between two \\textit{arbitrary} sampleable distributions without being constrained to a specific form as in the KL divergence, and without requiring expensive adversarial training.\nTo improve the representation, we use (1) a structural consistency term in the loss that encourages isometry between feature space and latent space and (2) a nonlinear variant of the SW distance which averages over random nonlinear shearing.\nThe effectiveness of the learned manifold encoding is best explored by traversing the latent space through interpolations along \\textit{geodesics} which generate samples that lie on the manifold and hence are advantageous compared to standard Euclidean interpolation.\nTo this end, we introduce a graph-based algorithm for interpolating along network-geodesics in latent space by maximizing the density of samples along the path while minimizing total energy. We use the 3D-spiral data to show that the prior does indeed encode the geometry underlying the data and to demonstrate the advantages of the network-algorithm for interpolation.\nAdditionally, we apply our framework to MNIST, and CelebA datasets, and show that outlier generations, latent representations, and geodesic interpolations are comparable to the state of the art.",
      "url": "https://openreview.net/forum?id=5L8XMh667qz",
      "year": 2021,
      "venue": "ICLR 2021",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "5L8XMh667qz",
      "track": "main",
      "status": "Reject",
      "keywords": "VAE;sliced Wasserstein distance;latent representation;interpolation;manifold embedding;geodesics;network algorithm",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 12.171001808720845,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.171001808720845,
      "combined_score": 0.0,
      "rank": 30
    },
    {
      "title": "Adam Reduces a Unique Form of Sharpness: Theoretical Insights Near the Minimizer Manifold",
      "authors": [
        "Xinghan Li",
        "Haodong Wen",
        "Kaifeng Lyu"
      ],
      "abstract": "Despite the popularity of Adam optimizer in practice, most theoretical analyses study SGD as a proxy and little is known about how the solutions found by Adam differ. In this paper, we show that Adam reduces a specific form of sharpness measure shaped by its adaptive updates, leading to qualitatively different solutions from SGD.\n    When the training loss is small, Adam wanders around the manifold of minimizers and takes semi-gradients to minimize this sharpness measure in an adaptive manner, a behavior we rigorously characterize via a continuous-time approximation using stochastic differential equations.\n    We further illustrate how this behavior differs from that of SGD in a well-studied setting: when training overparameterized models with label noise, SGD has been shown to minimize the trace of the Hessian matrix, $\\text{tr}(\\textbf{H})$, whereas we prove that Adam minimizes $\\text{tr}(\\text{diag}(\\textbf{H})^{1/2})$ instead. In solving sparse linear regression with diagonal linear networks, Adam provably achieves better sparsity and generalization than SGD due to this difference.\n    Finally, we note that our proof framework applies not only to Adam but also to a broad class of adaptive gradient methods, including but not limited to RMSProp, Adam-mini, and Adalayer. This provides a unified perspective for analyzing how adaptive optimizers reduce sharpness and may offer insights for future optimizer design.",
      "url": "https://openreview.net/forum?id=kCUDzyKQ7G",
      "year": 2025,
      "venue": "NIPS 2025",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "kCUDzyKQ7G",
      "track": "main",
      "status": "Poster",
      "keywords": "Adam;adaptive gradient methods;implicit bias;regularization;deep learning theory",
      "tldr": "",
      "primary_area": "theory",
      "similarity_score": 12.158096854609315,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.158096854609315,
      "combined_score": 0.0,
      "rank": 31
    },
    {
      "title": "Learning Globally Smooth Functions on Manifolds",
      "authors": [
        "Juan Cervino",
        "Luiz F. O. Chamon",
        "Benjamin David Haeffele",
        "Rene Vidal",
        "Alejandro Ribeiro"
      ],
      "abstract": "Smoothness and low dimensional structures play central roles in improving generalization and stability in learning and statistics. This work combines techniques from semi-infinite constrained learning and manifold regularization to learn representations that are globally smooth on a manifold. To do so, it shows that under typical conditions the problem of learning a Lipschitz continuous function on a manifold is equivalent to a dynamically weighted manifold regularization problem. This observation leads to a practical algorithm based on a weighted Laplacian penalty whose weights are adapted using stochastic gradient techniques. It is shown that under mild conditions, this method estimates the Lipschitz constant of the solution, learning a globally smooth solution as a byproduct. Experiments on real world data illustrate the advantages of the proposed method relative to existing alternatives. Our code is available at https://github.com/JuanCervino/smoothbench.",
      "url": "https://icml.cc/virtual/2023/poster/24397",
      "year": 2023,
      "venue": "ICML 2023",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=4bNGE4WSfJ",
      "citations": null,
      "categories": [],
      "id": "4bNGE4WSfJ",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 12.149153773767967,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 12.149153773767967,
      "combined_score": 0.0,
      "rank": 32
    },
    {
      "title": "The Momentum Persistence Effect: A New Theory for Why Soft Constraints Outperform Hard Projections",
      "authors": [
        "Khurram Khalil",
        "Ripan Kumar Kundu",
        "Khaza Anuarul Hoque"
      ],
      "abstract": "A persistent empirical puzzle in deep learning is why soft, penalty-based constraints often outperform their mathematically exact, hard-projected counterparts. While classical optimization theory provides elegant models, it fails to explain this phenomenon. This paper resolves the mystery by identifying a fundamental, theoretically unaccounted-for mechanism: the momentum persistence effect. We demonstrate that the classical theory assumes optimizer momentum resets after each projection, an assumption contradicted by standard implementations, such as Adam and SGD. Through controlled experiments on a tractable quadratic problem, we first show that the \\textit{``momentum reset\"} model fails catastrophically, under-predicting corruption magnitudes by orders of magnitude and misjudging scaling laws with respect to learning rate, projection frequency, and problem conditioning. We then isolate the cause through a crucial experiment: when momentum persists across projections, as in practice, the inherited optimizer state compounds corruption, leading to saturation at levels orders of magnitude higher than in memory-less cycles. Our corrected model accurately predicts this saturation and explains the observed super-linear scaling relationships. We further validate these principles in large-scale Transformer models using \\textit{Orthogonal Subspace Projection Attention (OSPA)}, confirming that momentum persistence has a significant impact on performance, particularly in high-noise, low-data scenarios. Our discovery reveals a critical blind spot in constrained optimization theory and provides key design principles for practitioners: prefer soft constraints when possible, and when hard projections are necessary, co-design them with optimizer choice to minimize momentum corruption effects.",
      "url": "https://openreview.net/forum?id=a0kq0tJwwn",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "a0kq0tJwwn",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Constrained Optimization;Deep Learning Theory;Optimization Dynamics;Momentum Methods;Orthogonal Constraints;Regularization;Stiefel Manifold",
      "tldr": "",
      "primary_area": "optimization",
      "similarity_score": 11.966275687901245,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.966275687901245,
      "combined_score": 0.0,
      "rank": 33
    },
    {
      "title": "Manifold Precis: An Annealing Technique for Diverse Sampling of Manifolds",
      "authors": [
        "Nitesh Shroff",
        "Pavan Turaga",
        "Rama Chellappa"
      ],
      "abstract": "In this paper, we consider the 'Precis' problem of sampling K representative yet diverse data points from a large dataset. This problem arises frequently in applications such as video and document summarization, exploratory data analysis, and pre-filtering. We formulate a general theory which encompasses not just traditional techniques devised for vector spaces, but also non-Euclidean manifolds, thereby enabling these techniques to shapes, human activities, textures and many other image and video based datasets. We propose intrinsic manifold measures for measuring the quality of a selection of points with respect to their representative power, and their diversity. We then propose efficient algorithms to optimize the cost function using a novel annealing-based iterative alternation algorithm. The proposed formulation is applicable to manifolds of known geometry as well as to manifolds whose geometry needs to be estimated from samples. Experimental results show the strength and generality of the proposed approach.",
      "url": "https://papers.nips.cc/paper_files/paper/2011/hash/d1f491a404d6854880943e5c3cd9ca25-Abstract.html",
      "year": 2011,
      "venue": "NIPS 2011",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2011/file/d1f491a404d6854880943e5c3cd9ca25-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "8ba6358e96",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.860893465744704,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.860893465744704,
      "combined_score": 0.0,
      "rank": 34
    },
    {
      "title": "Random matrix theory improved Fréchet mean of symmetric positive definite matrices",
      "authors": [
        "Florent Bouchard",
        "Ammar Mian",
        "Malik Tiomoko",
        "Guillaume Ginolhac",
        "Frederic Pascal"
      ],
      "abstract": "In this study, we consider the realm of covariance matrices in machine learning, particularly focusing on computing Fréchet means on the manifold of symmetric positive definite matrices, commonly referred to as Karcher or geometric means. Such means are leveraged in numerous machine learning tasks. Relying on advanced statistical tools, we introduce a random matrix theory based method that estimates Fréchet means, which is particularly beneficial when dealing with low sample support and a high number of matrices to average. Our experimental evaluation, involving both synthetic and real-world EEG and hyperspectral datasets, shows that we largely outperform state-of-the-art methods.",
      "url": "https://icml.cc/virtual/2024/poster/32828",
      "year": 2024,
      "venue": "ICML 2024",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=uQiFsBil3p",
      "citations": null,
      "categories": [],
      "id": "uQiFsBil3p",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.852463168964604,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.852463168964604,
      "combined_score": 0.0,
      "rank": 35
    },
    {
      "title": "Emergence of meta-stable clustering in mean-field transformer models",
      "authors": [
        "Giuseppe Bruno",
        "Federico Pasqualotto",
        "Andrea Agazzi"
      ],
      "abstract": "We model the evolution of tokens within a deep stack of Transformer layers as a continuous-time flow on the unit sphere, governed by a mean-field interacting particle system, building on the framework introduced in Geshkovski et al. (2023). Studying the corresponding mean-field Partial Differential Equation (PDE), which can be interpreted as a Wasserstein gradient flow, in this paper we provide a mathematical investigation of the long-term behavior of this system, with a particular focus on the emergence and persistence of meta-stable phases and clustering phenomena, key elements in applications like next-token prediction. More specifically, we perform a perturbative analysis of the mean-field PDE around the iid uniform initialization and prove that, in the limit of large number of tokens, the model remains close to a meta-stable manifold of solutions with a given structure (e.g., periodicity). Further, the structure characterizing the meta-stable manifold is explicitly identified, as a function of the inverse temperature parameter of the model, by the index maximizing a certain rescaling of Gegenbauer polynomials.",
      "url": "https://iclr.cc/virtual/2025/poster/28944",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=eBS3dQQ8GV",
      "citations": null,
      "categories": [],
      "id": "eBS3dQQ8GV",
      "track": "main",
      "status": "Oral",
      "keywords": "Mean-field limits;Transformers;Meta-stability;Clustering",
      "tldr": "",
      "primary_area": "foundation or frontier models, including LLMs",
      "similarity_score": 11.83438124735372,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.83438124735372,
      "combined_score": 0.0,
      "rank": 36
    },
    {
      "title": "Option Discovery in the Absence of Rewards with Manifold Analysis",
      "authors": [
        "Amitay Bar",
        "Ronen Talmon",
        "Ron Meir"
      ],
      "abstract": "Options have been shown to be an effective tool in reinforcement learning, facilitating improved exploration and learning. In this paper, we present an approach based on spectral graph theory and derive an algorithm that systematically discovers options without access to a specific reward or task assignment. As opposed to the common practice used in previous methods, our algorithm makes full use of the spectrum of the graph Laplacian. Incorporating modes associated with higher graph frequencies unravels domain subtleties, which are shown to be useful for option discovery. Using geometric and manifold-based analysis, we present a theoretical justification for the algorithm. In addition, we showcase its performance in several domains, demonstrating clear improvements compared to competing methods.",
      "url": "https://icml.cc/virtual/2020/poster/6413",
      "year": 2020,
      "venue": "ICML 2020",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "http://proceedings.mlr.press/v119/bar20a/bar20a.pdf",
      "citations": null,
      "categories": [],
      "id": "6413",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.796019071648487,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.796019071648487,
      "combined_score": 0.0,
      "rank": 37
    },
    {
      "title": "The Sparse Manifold Transform",
      "authors": [
        "Yubei Chen",
        "Dylan Paiton",
        "Bruno Olshausen"
      ],
      "abstract": "We present a signal representation framework called the sparse manifold transform that combines key ideas from sparse coding, manifold learning, and slow feature analysis. It turns non-linear transformations in the primary sensory signal space into linear interpolations in a representational embedding space while maintaining approximate invertibility. The sparse manifold transform is an unsupervised and generative framework that explicitly and simultaneously models the sparse discreteness and low-dimensional manifold structure found in natural scenes. When stacked, it also models hierarchical composition. We provide a theoretical description of the transform and demonstrate properties of the learned representation on both synthetic data and natural videos.",
      "url": "https://nips.cc/virtual/2018/poster/11994",
      "year": 2018,
      "venue": "NIPS 2018",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2018/file/8e19a39c36b8e5e3afd2a3b2692aea96-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "11994",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.779456807981608,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.779456807981608,
      "combined_score": 0.0,
      "rank": 38
    },
    {
      "title": "Density Constrained Reinforcement Learning",
      "authors": [
        "Zengyi Qin",
        "Yuxiao Chen",
        "Chuchu Fan"
      ],
      "abstract": "We study constrained reinforcement learning (CRL) from a novel perspective by setting constraints directly on state density functions, rather than the value functions considered by previous works. State density has a clear physical and mathematical interpretation, and is able to express a wide variety of constraints such as resource limits and safety requirements. Density constraints can also avoid the time-consuming process of designing and tuning cost functions required by value function-based constraints to encode system specifications. We leverage the duality between density functions and Q functions to develop an effective algorithm to solve the density constrained RL problem optimally and the constrains are guaranteed to be satisfied. We prove that the proposed algorithm converges to a near-optimal solution with a bounded error even when the policy update is imperfect. We use a set of comprehensive experiments to demonstrate the advantages of our approach over state-of-the-art CRL methods, with a wide range of density constrained tasks as well as standard CRL benchmarks such as Safety-Gym.",
      "url": "https://icml.cc/virtual/2021/poster/8985",
      "year": 2021,
      "venue": "ICML 2021",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "http://proceedings.mlr.press/v139/qin21a/qin21a.pdf",
      "citations": null,
      "categories": [],
      "id": "8985",
      "track": "main",
      "status": "Spotlight",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.755163795703432,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.755163795703432,
      "combined_score": 0.0,
      "rank": 39
    },
    {
      "title": "NaturalProver: Grounded Mathematical Proof Generation with Language Models",
      "authors": [
        "Sean Welleck",
        "Jiacheng Liu",
        "Ximing Lu",
        "Hannaneh Hajishirzi",
        "Yejin Choi"
      ],
      "abstract": "Theorem proving in natural mathematical language – the mixture of symbolic and natural language used by humans – plays a central role in mathematical advances and education, and tests aspects of reasoning that are core to intelligence. Yet it has remained underexplored with modern generative models. We study large-scale language models on two new generation tasks: suggesting the next step in a mathematical proof, and full proof generation. We develop NaturalProver, a language model that generates proofs by conditioning on background references (e.g. theorems and definitions that are either retrieved or human-provided), and optionally enforces their presence with constrained decoding. On theorems from the NaturalProofs benchmark, NaturalProver improves the quality of next-step suggestions and generated proofs over fine-tuned GPT-3, according to human evaluations from university-level mathematics students. NaturalProver is capable of proving some theorems that require short (2-6 step) proofs, and providing next-step suggestions that are rated as correct and useful over 40% of the time, which is to our knowledge the first demonstration of these capabilities using neural language models.",
      "url": "https://nips.cc/virtual/2022/poster/53914",
      "year": 2022,
      "venue": "NIPS 2022",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=rhdfTOiXBng",
      "citations": null,
      "categories": [],
      "id": "rhdfTOiXBng",
      "track": "main",
      "status": "Accept",
      "keywords": "language modeling;reasoning;neural theorem proving",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.695810896356589,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.695810896356589,
      "combined_score": 0.0,
      "rank": 40
    },
    {
      "title": "Noisy Feature Mixup",
      "authors": [
        "Soon Hoe Lim",
        "N. Benjamin Erichson",
        "Francisco Utrera",
        "Winnie Xu",
        "Michael W. Mahoney"
      ],
      "abstract": "We introduce Noisy Feature Mixup (NFM), an inexpensive yet effective method for data augmentation that combines the best of interpolation based training and noise injection schemes. Rather than training with convex combinations of pairs of examples and their labels, we use noise-perturbed convex combinations of pairs of data points in both input and feature space. This method includes mixup and manifold mixup as special cases, but it has additional advantages, including better smoothing of decision boundaries and enabling improved model robustness. We provide theory to understand this as well as the implicit regularization effects of NFM. Our theory is supported by empirical results, demonstrating the advantage of NFM, as compared to mixup and manifold mixup. We show that residual networks and vision transformers trained with NFM have favorable trade-offs between predictive accuracy on clean data and robustness with respect to various types of data perturbation across a range of computer vision benchmark datasets.",
      "url": "https://iclr.cc/virtual/2022/poster/6227",
      "year": 2022,
      "venue": "ICLR 2022",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=vJb4I2ANmy",
      "citations": null,
      "categories": [],
      "id": "vJb4I2ANmy",
      "track": "main",
      "status": "Poster",
      "keywords": "Data augmentation;implicit regularization;mixup;noise injection;model robustness",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.65734517741608,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.65734517741608,
      "combined_score": 0.0,
      "rank": 41
    },
    {
      "title": "Finding One Missing Puzzle of Contextual Word Embedding: Representing Contexts as Manifold",
      "authors": [
        "Hailin Hu",
        "Rong Yao",
        "Cheng LI"
      ],
      "abstract": "The current understanding of contextual word embedding interprets the representation by associating each token to a vector that is dynamically modulated by the context. However, this “token-centric” understanding does not explain how a model represents context itself, leading to a lack of characterization from such a perspective. In this work, to establish a rigorous definition of “context representation”, we formalize this intuition using a category theory framework, which indicates the necessity of including the information from both tokens and how transitions happen among different tokens in a given context. As a practical instantiation of our theoretical understanding, we also show how to leverage a manifold learning method to characterize how a representation model (i.e., BERT) encodes different contexts and how a representation of context changes when going through different components such as attention and FFN. We hope this novel theoretic perspective sheds light on the further improvements in Transformer-based language representation models.",
      "url": "https://openreview.net/forum?id=m7zsaLt1Sab",
      "year": 2022,
      "venue": "ICLR 2022",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "m7zsaLt1Sab",
      "track": "main",
      "status": "Reject",
      "keywords": "Contextual Word Embedding;Category Theory;Manifold",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.656116024921712,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.656116024921712,
      "combined_score": 0.0,
      "rank": 42
    },
    {
      "title": "TopoAlign: A Framework for Aligning Code to Math via Topological Decomposition",
      "authors": [],
      "abstract": "Large Language Models (LLMs) excel at both informal and formal (e.g. Lean 4) mathematical reasoning but still struggle with autoformalisation, the task of transforming informal into formal mathematical statements. Autoformalisation helps pair the informal reasoning of LLMs with formal proof assistants which enable machine-verifiable generation and mitigate hallucinations. Yet, the performance of current Math LLMs is constrained by the scarcity of large-scale corpora, particularly those containing pairs of informal and formal statements. Although current models are trained to generate code from natural language instructions, structural and syntactic differences between these and formal mathematics limit effective transfer learning. We propose TopoAlign, a framework that unlocks widely available code repositories as training resources for Math LLMs. TopoAlign decomposes code into docstrings, main functions, and dependency functions, and reassembles these components into analogues that structurally mirror formal statements. This produces structurally aligned code data that can be used for training Math LLMs without requiring additional human annotation.  We train two state-of-the-art models, DeepSeek-Math and Herald, and evaluate them on the minif2f, Putnam, and ProofNet benchmarks. TopoAlign provides substantial gains for DeepSeek-Math, improving performance by 17.77% on BEq@10 and 68.82% on typecheck@10. Despite introducing no new mathematical knowledge, our framework achieves gains of 0.12% and 1.09% for Herald on BEq@10 and typecheck@10, respectively, demonstrating that training on aligned code data is beneficial even for specialized models.",
      "url": "https://openreview.net/forum?id=OEPP8T0zUX",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "OEPP8T0zUX",
      "track": "main",
      "status": "Active",
      "keywords": "autoformalisation;formal reasoning;code data",
      "tldr": "",
      "primary_area": "neurosymbolic & hybrid AI systems (physics-informed, logic & formal reasoning, etc.)",
      "similarity_score": 11.652344668380234,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.652344668380234,
      "combined_score": 0.0,
      "rank": 43
    },
    {
      "title": "Matrix Manifold Neural Networks++",
      "authors": [
        "Xuan Son Nguyen",
        "Shuo Yang",
        "Aymeric Histace"
      ],
      "abstract": "Deep neural networks (DNNs) on Riemannian manifolds have garnered increasing interest in various applied areas. For instance, DNNs on spherical and hyperbolic manifolds have been designed to solve a wide range of computer vision and nature language processing tasks. One of the key factors that contribute to the success of these networks is that spherical and hyperbolic manifolds have the rich algebraic structures of gyrogroups and gyrovector spaces. This enables principled and effective generalizations of the most successful DNNs to these manifolds. Recently, some works have shown that many concepts in the theory of gyrogroups and gyrovector spaces can also be generalized to matrix manifolds such as Symmetric Positive Definite (SPD) and Grassmann manifolds. As a result, some building blocks for SPD and Grassmann neural networks, e.g., isometric models and multinomial logistic regression (MLR) can be derived in a way that is fully analogous to their spherical and hyperbolic counterparts. Building upon these works, in this paper, we design fully-connected (FC) and convolutional layers for SPD neural networks. We also develop MLR on Symmetric Positive Semi-definite (SPSD) manifolds, and propose a method for performing backpropagation with the Grassmann logarithmic map in the projector perspective. We demonstrate the effectiveness of the proposed approach in the human action recognition and node classification tasks.",
      "url": "https://iclr.cc/virtual/2024/poster/19532",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=30aSE3FB3L",
      "citations": null,
      "categories": [],
      "id": "30aSE3FB3L",
      "track": "main",
      "status": "Poster",
      "keywords": "manifold learning;representation learning;gyrovector spaces;deep learning",
      "tldr": "",
      "primary_area": "representation learning for computer vision, audio, language, and other modalities",
      "similarity_score": 11.635716522307614,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.635716522307614,
      "combined_score": 0.0,
      "rank": 44
    },
    {
      "title": "Sparse Quadratic Optimisation over the Stiefel Manifold with Application to Permutation Synchronisation",
      "authors": [
        "Florian Bernard",
        "Daniel Cremers",
        "Anders Johan Thunberg"
      ],
      "abstract": "We address the non-convex optimisation problem of finding a sparse matrix on the Stiefel manifold (matrices with mutually orthogonal columns of unit length) that maximises (or minimises) a quadratic objective function. Optimisation problems on the Stiefel manifold occur for example in spectral relaxations of various combinatorial problems, such as graph matching, clustering, or permutation synchronisation. Although sparsity is a desirable property in such settings, it is mostly neglected in spectral formulations since existing solvers, e.g. based on eigenvalue decomposition, are unable to account for sparsity while at the same time maintaining global optimality guarantees. We fill this gap and propose a simple yet effective sparsity-promoting modification of the Orthogonal Iteration algorithm for finding the dominant eigenspace of a matrix. By doing so, we can guarantee that our method finds a Stiefel matrix that is globally optimal with respect to the quadratic objective function, while in addition being sparse. As a motivating application we consider the task of permutation synchronisation, which can be understood as a constrained clustering problem that has particular relevance for matching multiple images or 3D shapes in computer vision, computer graphics, and beyond. We demonstrate that the proposed approach outperforms previous methods in this domain.",
      "url": "https://nips.cc/virtual/2021/poster/28158",
      "year": 2021,
      "venue": "NIPS 2021",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=sl_0rQmHxQk",
      "citations": null,
      "categories": [],
      "id": "sl_0rQmHxQk",
      "track": "main",
      "status": "Poster",
      "keywords": "Stiefel manifold;quadratic optimisation;permutation synchronisation;sparsity;multi-matching;correspondence problems;manifold optimisation;QR decomposition;orthogonal iteration algorithm",
      "tldr": "A method for finding a globally optimal solution of a quadratic objective function over the Stiefel manifold that is sparse.",
      "primary_area": "",
      "similarity_score": 11.630622989068623,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.630622989068623,
      "combined_score": 0.0,
      "rank": 45
    },
    {
      "title": "Riemannian Manifold Learning for Stackelberg Games with Neural Flow Representations",
      "authors": [
        "Larkin Liu",
        "Yutong Chao",
        "Jalal Etesami",
        "Kashif Rasul"
      ],
      "abstract": "We present a novel framework for online learning in Stackelberg general-sum games, where two agents, the leader and follower, engage in sequential turn-based interactions. At the core of this approach is a learned diffeomorphism that maps the joint action space to a smooth Riemannian manifold, referred to as the $\\textit{Stackelberg manifold}$. This mapping, facilitated by neural normalizing flows, ensures the formation of tractable isoplanar subspaces, enabling efficient techniques for online learning. By assuming linearity between the agents' reward functions on the $\\textit{Stackelberg manifold}$, our construct allows the application of standard bandit algorithms. We then provide a rigorous theoretical basis for regret minimization on convex manifolds and establish finite-time bounds on simple regret for learning Stackelberg equilibria. This integration of manifold learning into game theory uncovers a previously unrecognized potential for neural normalizing flows as an effective tool for multi-agent learning. We present empirical results demonstrating the effectiveness of our approach compared to standard baselines, with applications spanning domains such as cybersecurity and economic supply chain optimization.",
      "url": "https://openreview.net/forum?id=eAFNJk63KE",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "eAFNJk63KE",
      "track": "main",
      "status": "Reject",
      "keywords": "Neural Normalizing Flows;Stackelberg Games;Riemannian Manifolds",
      "tldr": "",
      "primary_area": "learning on graphs and other geometries & topologies",
      "similarity_score": 11.599912084087297,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.599912084087297,
      "combined_score": 0.0,
      "rank": 46
    },
    {
      "title": "GeoMind: A Geometric Neural Network of State Space Model for Understanding Brain Dynamics on Riemannian Manifold",
      "authors": [
        "Tingting Dan",
        "Jiaqi Ding",
        "Guorong Wu"
      ],
      "abstract": "State space model (SSM) is a powerful tool in neuroscience field to characterize the dynamic nature of brain functions by elucidating the mechanism of how brain system transits between brain states and how underlying states give rise to the observed neural activities. Although tremendous efforts have been made to lend the power of deep learning and mathematical insight of SSM in various functional neuroimaging studies, current state-of-the-art methods lack a holistic view of brain state evolution as a self-organized dynamical system where each part of the brain is functionally inter-connected. Since the topological co-activation of functional fluctuations exhibits an intrinsic geometric pattern (symmetric and positive definite, or SPD) on the Riemannian manifold, the call for understanding how a selective set of functional connectivities in the brain supports diverse behavior and cognition emerges a new machine learning scenario of manifold-based SSM for large-scale functional neuroimages. To that end, we propose a geometric neural networks, coined *GeoMind*, designed to uncover evolving brain states by tracking the trajectory of functional dynamics on a high-dimensional Riemannian manifold of SPD matrices. Our *GeoMind* demonstrates promising results in identifying specific brain states based on task-based functional Magnetic Resonance Imaging (fMRI) data, as well as in diseases early diagnosis for Alzheimer's disease, Parkinson's disease and Autism. These results highlight the applicability of the proposed *GeoMind* in neuroscience research. Furthermore, to assess the generalization capabilities of our model, we applied it to the domain of human action recognition (HAR), achieving promising performance on three benchmark datasets (UTKinect, Florence and HDM05). This demonstrates the scalability and robustness of the proposed geometry deep model of SSM in capturing complex spatio-temporal dynamics across diverse fields.",
      "url": "https://openreview.net/forum?id=YZdc7mTq7I",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "YZdc7mTq7I",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Geometric deep learning;state space model;brain dynamics;Riemannian Manifold",
      "tldr": "",
      "primary_area": "applications to neuroscience & cognitive science",
      "similarity_score": 11.431538685155795,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.431538685155795,
      "combined_score": 0.0,
      "rank": 47
    },
    {
      "title": "Effects of Data Geometry in Early Deep Learning",
      "authors": [
        "Saket Tiwari",
        "George Konidaris"
      ],
      "abstract": "Deep neural networks can approximate functions on different types of data, from images to graphs, with varied underlying structure.This underlying structure can be viewed as the geometry of the data manifold. By extending recent advances in the theoretical understanding of neural networks, we study how a randomly initialized neural network with piecewise linear activation splits the data manifold into regions where the neural network behaves as a linear function. We derive bounds on the number of linear regions and the distance to boundaries of these linear regions on the data manifold. This leads to insights into the expressivity of randomly initialized deep neural networks on non-Euclidean data sets. We empirically corroborate our theoretical results using a toy supervised learning problem. Our experiments demonstrate that number of linear regions varies across manifolds and how our results hold upon changing neural network architectures. We further demonstrate how the complexity of linear regions changes on the low dimensional manifold of images as training progresses, using the MetFaces dataset.",
      "url": "https://openreview.net/forum?id=vKMVrqvXbXu",
      "year": 2022,
      "venue": "ICLR 2022",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "vKMVrqvXbXu",
      "track": "main",
      "status": "Reject",
      "keywords": "Deep learning;geometry;manifolds;deep learning theory",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.397485907525542,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.397485907525542,
      "combined_score": 0.0,
      "rank": 48
    },
    {
      "title": "On Deep Generative Models for Approximation and Estimation of Distributions on Manifolds",
      "authors": [
        "Biraj Dahal",
        "Alexander Havrilla",
        "Minshuo Chen",
        "Tuo Zhao",
        "Wenjing Liao"
      ],
      "abstract": "Deep generative models have experienced great empirical successes in distribution learning. Many existing experiments have demonstrated that deep generative networks can efficiently generate high-dimensional complex data from a low-dimensional easy-to-sample distribution. However, this phenomenon can not be justified by existing theories. The widely held manifold hypothesis speculates that real-world data sets, such as natural images and signals, exhibit low-dimensional geometric structures. In this paper, we take such low-dimensional data structures into consideration by assuming that data distributions are supported on a low-dimensional manifold. We prove approximation and estimation theories of deep generative networks for estimating distributions on a low-dimensional manifold under the Wasserstein-1 loss. We show that the Wasserstein-1 loss converges to zero at a fast rate depending on the intrinsic dimension instead of the ambient data dimension. Our theory leverages the low-dimensional geometric structures in data sets and justifies the practical power of deep generative models. We require no smoothness assumptions on the data distribution which is desirable in practice.",
      "url": "https://nips.cc/virtual/2022/poster/53839",
      "year": 2022,
      "venue": "NIPS 2022",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=4n1PS9WvdYv",
      "citations": null,
      "categories": [],
      "id": "4n1PS9WvdYv",
      "track": "main",
      "status": "Accept",
      "keywords": "Deep generative models;distribution estimation;low-dimensional manifold",
      "tldr": "We prove approximation and statistical estimation theories of deep generative models for distribution learning when the distribution is supported on a low-dimensional manifold.",
      "primary_area": "",
      "similarity_score": 11.383968544283391,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.383968544283391,
      "combined_score": 0.0,
      "rank": 49
    },
    {
      "title": "A solvable model of learning generative diffusion: theory and insights",
      "authors": [
        "Hugo Cui",
        "Cengiz Pehlevan",
        "Yue M. Lu"
      ],
      "abstract": "In this manuscript, we analyze a solvable model of flow or diffusion-based generative model. We consider the problem of learning a model parametrized by a two-layer auto-encoder, trained with online stochastic gradient descent, on a high-dimensional target density with an underlying low-dimensional manifold structure. We derive a tight asymptotic characterization of low-dimensional projections of the distribution of samples generated by the learned model, ascertaining in particular its dependence on the number of training samples. Building on this analysis, we discuss how mode collapse can arise, and lead to model collapse when the generative model is re-trained on generated synthetic data.",
      "url": "https://openreview.net/forum?id=5b5wZg6Zeo",
      "year": 2025,
      "venue": "NIPS 2025",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "5b5wZg6Zeo",
      "track": "main",
      "status": "Poster",
      "keywords": "high-dimensional asymptotics;statistical physics;diffusion model",
      "tldr": "",
      "primary_area": "theory",
      "similarity_score": 11.35180424304155,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.35180424304155,
      "combined_score": 0.0,
      "rank": 50
    },
    {
      "title": "Retraction-free optimization over the Stiefel manifold with application to the LoRA fine-tuning",
      "authors": [
        "Yuan Zhang",
        "Jiang Hu",
        "Jiaxi Cui",
        "Lin Lin",
        "Zaiwen Wen",
        "Quanzheng Li"
      ],
      "abstract": "Optimization over the Stiefel manifold has played a significant role in various machine learning tasks. Many existing algorithms either use the retraction operator to keep each iterate staying on the manifold, or solve an unconstrained quadratic penalized problem. The retraction operator in the former corresponds to orthonormalization of matrices and can be computationally costly for large-scale matrices. The latter approach usually equips with an unknown large penalty parameter. To address the above issues, we propose a retraction-free and penalty parameter-free algorithm, which lands on the manifold. Moreover, our convergence theory allows using constant step size, which improve the result of converging to a neighborhood in \\citep{ablin2022fast}.\n  A key component of the analysis is the convex-like property of the quadratic penalty of the Stiefel manifold, which enables us to explicitly characterize the constant penalty parameter. As an application, we introduce a new algorithm, Manifold-LoRA, which employs the landing technique and a carefully designed step size strategy to accelerate low-rank adaptation (LoRA) in fine-tuning large language models. Numerical experiments on the benchmark datasets demonstrate the efficiency of our proposed method.",
      "url": "https://openreview.net/forum?id=c2OtbtZXFC",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "c2OtbtZXFC",
      "track": "main",
      "status": "Withdraw",
      "keywords": "landing;manifold;fine-tuning;LoRA",
      "tldr": "",
      "primary_area": "optimization",
      "similarity_score": 11.347613250889898,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.347613250889898,
      "combined_score": 0.0,
      "rank": 51
    },
    {
      "title": "BoostStep: Boosting Mathematical Capability of Large Language Models via Step-aligned In Context Learning",
      "authors": [],
      "abstract": "Large language models (LLMs) have demonstrated impressive ability in solving complex mathematical problems with multi-step reasoning and can be further enhanced with well-designed in-context learning (ICL) examples. However, this potential is often constrained by two major challenges in ICL: granularity mismatch and irrelevant information.\nWe observe that while LLMs excel at decomposing mathematical problems, they often struggle with reasoning errors in fine-grained steps. Moreover, ICL examples retrieved at the question level may omit critical steps or even mislead the model with irrelevant details.\nTo address this issue, we propose BoostStep, a method that enhances reasoning accuracy through step-aligned ICL, a novel mechanism that carefully aligns retrieved reference steps with the corresponding reasoning steps. Additionally, BoostStep incorporates an effective \"first-try\" strategy to retrieve for exemplars highly relevant to the current state of reasoning.\nBoostStep is a flexible and powerful method that integrates seamlessly with chain-of-thought (CoT) and tree search algorithms, refining both candidate selection and decision-making. Empirical results show that BoostStep improves GPT-4o’s CoT performance by 4.6\\% across mathematical benchmarks, significantly surpassing traditional few-shot learning's 1.2\\%. Moreover, it can achieve an additional 7.5\\% gain combined with tree search. Surprisingly, it enhances state-of-the-art LLMs to solve challenging math problems using simpler examples. It improves DeepSeek-R1-671B and Qwen3-235B’s performance on AIME by 2.2\\% and 5.0\\% respectively, leveraging simple examples only from the MATH dataset.",
      "url": "https://openreview.net/forum?id=TXJ7vLgOS4",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "TXJ7vLgOS4",
      "track": "main",
      "status": "Active",
      "keywords": "Mathematical Reasoning;Large Language Models;In-context Learning",
      "tldr": "",
      "primary_area": "foundation or frontier models, including LLMs",
      "similarity_score": 11.327626756765737,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.327626756765737,
      "combined_score": 0.0,
      "rank": 52
    },
    {
      "title": "Learning Multiple Tasks using Manifold Regularization",
      "authors": [
        "Arvind Agarwal",
        "Samuel Gerber",
        "Hal Daume"
      ],
      "abstract": "We present a novel method for multitask learning (MTL) based on {\\it manifold regularization}: assume that all task parameters lie on a manifold. This is the generalization of a common assumption made in the existing literature: task parameters share a common {\\it linear} subspace. One proposed method uses the projection distance from the manifold to regularize the task parameters. The manifold structure and the task parameters are learned using an alternating optimization framework. When the manifold structure is fixed, our method decomposes across tasks which can be learnt independently. An approximation of the manifold regularization scheme is presented that preserves the convexity of the single task learning problem, and makes the proposed MTL framework efficient and easy to implement. We show the efficacy of our method on several datasets.",
      "url": "https://papers.nips.cc/paper_files/paper/2010/hash/2cbca44843a864533ec05b321ae1f9d1-Abstract.html",
      "year": 2010,
      "venue": "NIPS 2010",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2010/file/2cbca44843a864533ec05b321ae1f9d1-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "e979a4d03a",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.324825876000121,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.324825876000121,
      "combined_score": 0.0,
      "rank": 53
    },
    {
      "title": "A Theory of Transfer-Based Black-Box Attacks: Explanation and Implications",
      "authors": [
        "Yanbo Chen",
        "Weiwei Liu"
      ],
      "abstract": "Transfer-based attacks are a practical method of black-box adversarial attacks, in which the attacker aims to craft adversarial examples from a source (surrogate) model that is transferable to the target model. A wide range of empirical works has tried to explain the transferability of adversarial examples from different angles. However, these works only provide ad hoc explanations without quantitative analyses. The theory behind transfer-based attacks remains a mystery.\nThis paper studies transfer-based attacks under a unified theoretical framework. We propose an explanatory model, called the manifold attack model, that formalizes popular beliefs and explains the existing empirical results. Our model explains why adversarial examples are transferable even when the source model is inaccurate. Moreover, our model implies that the existence of transferable adversarial examples depends on the “curvature” of the data manifold, which quantitatively explains why the success rates of transfer-based attacks are hard to improve. We also discuss the expressive power and the possible extensions of our model in general applications.",
      "url": "https://nips.cc/virtual/2023/poster/72434",
      "year": 2023,
      "venue": "NIPS 2023",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=CJY7NEXVwC",
      "citations": null,
      "categories": [],
      "id": "CJY7NEXVwC",
      "track": "main",
      "status": "Poster",
      "keywords": "Learning Theory",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.306819226984135,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.306819226984135,
      "combined_score": 0.0,
      "rank": 54
    },
    {
      "title": "CodePlot-CoT: Mathematical Visual Reasoning by Thinking with Code-Driven Images",
      "authors": [
        "Chengqi Duan",
        "Kaiyue Sun",
        "Rongyao Fang",
        "Manyuan Zhang",
        "Yan Feng",
        "Ying Luo",
        "Yufang Liu",
        "Ke Wang",
        "Peng Pei",
        "Xunliang Cai"
      ],
      "abstract": "Recent advances in Large Language Models (LLMs) and Vision Language Models (VLMs) have shown significant progress in mathematical reasoning, yet they still face a critical bottleneck with problems requiring visual assistance, such as drawing auxiliary lines or plotting functions to solve the problems. Most LLMs and VLMs are constrained to text-only reasoning chains, while multimodal unified models that can generate interleaved text and images lack the necessary precision and controllability for such tasks. To address this, we propose CodePlot-CoT, a code-driven Chain-of-Thought paradigm for \"thinking with images\" in mathematics. Our approach leverages the VLM to generate text reasoning as well as executable plotting code, which is then rendered into images as \"visual thought\", to solve mathematical problems. To achieve this, we first construct Math-VR, the first large-scale, bilingual dataset and benchmark for Mathematics problems with Visual Reasoning, comprising 178K samples. Second, to create high-quality training data, we develop a state-of-the-art image-to-code converter specialized for parsing complex mathematical figures into codes. Finally, using these training data, we train the CodePlot-CoT model for solving mathematical problems. Experimental results show that our model achieves up to 21% increase over base model on our new benchmark, fully validating the efficacy of our proposed code-driven reasoning paradigm. Our work opens a new direction for multimodal mathematical reasoning and provides the community with the first large-scale dataset, comprehensive benchmark, and strong approach for such problems.",
      "url": "https://openreview.net/forum?id=zJaqyxO7K7",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "zJaqyxO7K7",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Multimodal Large Language Models;Mathematical Reasoning;Thinking with Images;Multimodal Benchmark",
      "tldr": "",
      "primary_area": "applications to computer vision, audio, language, and other modalities",
      "similarity_score": 11.30635829148358,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.30635829148358,
      "combined_score": 0.0,
      "rank": 55
    },
    {
      "title": "BAST: Bayesian Additive Regression Spanning Trees for Complex Constrained Domain",
      "authors": [
        "Zhao Tang Luo",
        "Huiyan Sang",
        "Bani Mallick"
      ],
      "abstract": "Nonparametric regression on complex domains has been a challenging task as most existing methods, such as ensemble models based on binary decision trees, are not designed to account for intrinsic geometries and domain boundaries. This article proposes a Bayesian additive regression spanning trees (BAST) model for nonparametric regression on manifolds, with an emphasis on complex constrained domains or irregularly shaped spaces embedded in Euclidean spaces. Our model is built upon a random spanning tree manifold partition model as each weak learner, which is capable of capturing any irregularly shaped spatially contiguous partitions while respecting intrinsic geometries and domain boundary constraints. Utilizing many nice properties of spanning tree structures, we design an efficient Bayesian inference algorithm. Equipped with a soft prediction scheme, BAST is demonstrated to significantly outperform other competing methods in simulation experiments and in an application to the chlorophyll data in Aral Sea, due to its strong local adaptivity to different levels of smoothness. ",
      "url": "https://nips.cc/virtual/2021/poster/28133",
      "year": 2021,
      "venue": "NIPS 2021",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=Yw7ZNeDVpBS",
      "citations": null,
      "categories": [],
      "id": "Yw7ZNeDVpBS",
      "track": "main",
      "status": "Poster",
      "keywords": "Bayesian nonparametric regression;Constrained domain;Ensemble learning;Manifold;Random spanning trees",
      "tldr": "This paper proposes a Bayesian nonparametric regression model on manifolds via additive random spanning tree partitions that adapts to different smoothness levels while respecting intrinsic geometries.",
      "primary_area": "",
      "similarity_score": 11.250162259727773,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.250162259727773,
      "combined_score": 0.0,
      "rank": 56
    },
    {
      "title": "A unifying framework for vector-valued manifold regularization and multi-view learning",
      "authors": [
        "Minh Hà Quang",
        "Loris Bazzani",
        "Vittorio Murino"
      ],
      "abstract": "This paper presents a general vector-valued reproducing kernel Hilbert spaces (RKHS) formulation for the problem of learning an unknown functional dependency  between a structured input space and a structured output space, in the Semi-Supervised Learning setting. Our formulation includes as special cases Vector-valued Manifold Regularization and Multi-view Learning, thus provides in particular a unifying framework linking these two important learning approaches.  In the case of least square loss function, we provide a closed form solution with an efficient implementation. Numerical experiments on challenging multi-class categorization problems show that our multi-view learning formulation achieves results which are comparable with state of the art and are significantly better than single-view learning.",
      "url": "https://proceedings.mlr.press/v28/haquang13.html",
      "year": 2013,
      "venue": "ICML 2013",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "http://proceedings.mlr.press/v28/haquang13.pdf",
      "citations": null,
      "categories": [],
      "id": "083947b5c2",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.23087740891812,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.23087740891812,
      "combined_score": 0.0,
      "rank": 57
    },
    {
      "title": "A Symbolic Framework for Evaluating Mathematical Reasoning with Transformers",
      "authors": [
        "Jordan Meadows",
        "Marco Valentino",
        "Damien Teney",
        "Andre Freitas"
      ],
      "abstract": "This paper proposes a methodology for generating synthetic mathematical derivations via a computer algebra system to evaluate the generalisability of Transformers in symbolic and quantitative reasoning problems, and provides a general framework for building large-scale and high-quality benchmarks in the mathematical domain. In the context of classification tasks involving multi-step annotated derivations (spanning 18 mathematical operators), we leverage the framework to compare the mathematical capabilities of GPT-4, GPT-3.5, and a canon of fine-tuned BERT models, exploring the relationship between specific operators and generalisation failure. Surprisingly, the average in-distribution performance of BERT models surpasses GPT-3.5, and rivals GPT-4, yet simple data perturbations reduce BERT scores by up to 80 F1 points. The results suggest that the in-distribution performance and generalisability of smaller open-source models may potentially rival GPT in narrow mathematical domains by incorporating appropriately structured discourse-level relations during training, and highlight a shared weakness between BERT and GPT involving a relative inability to decode dependency relations involving indirect references to mathematical entities. We release the data generation framework along with all the resulting datasets and fine-tuned models\\footnote{\\url{https://github.com/anonymous/TBA}}.",
      "url": "https://openreview.net/forum?id=7n8RzGQKnR",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "7n8RzGQKnR",
      "track": "main",
      "status": "Withdraw",
      "keywords": "mathematical reasoning;generalisation;gpt;bert;sequence classification;synthetic data;fine-tuning;few-shot learning",
      "tldr": "",
      "primary_area": "datasets and benchmarks",
      "similarity_score": 11.23065567064537,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.23065567064537,
      "combined_score": 0.0,
      "rank": 58
    },
    {
      "title": "Mitigating Overthinking in Large Reasoning Models via Manifold Steering",
      "authors": [
        "Yao Huang",
        "Huanran Chen",
        "Shouwei Ruan",
        "Yichi Zhang",
        "Xingxing Wei",
        "Yinpeng Dong"
      ],
      "abstract": "Recent advances in Large Reasoning Models (LRMs) have demonstrated remarkable capabilities in solving complex tasks such as mathematics and coding. However, these models frequently exhibit a phenomenon known as *overthinking* during inference, characterized by excessive validation loops and redundant deliberation, leading to substantial computational overheads. In this paper, we aim to mitigate overthinking by investigating the underlying mechanisms from the perspective of mechanistic interpretability. We first showcase that the tendency of overthinking can be effectively captured by a single direction in the model's activation space and the issue can be eased by intervening the activations along this direction. However, this efficacy soon reaches a plateau and even deteriorates as the intervention strength increases. We therefore systematically explore the activation space and find that the overthinking phenomenon is actually tied to a low-dimensional manifold, which indicates that the limited effect stems from the noises introduced by the high-dimensional steering direction. Based on this insight, we propose **Manifold Steering**, a novel approach that elegantly projects the steering direction onto the low-dimensional activation manifold given the theoretical approximation of the interference noise. Extensive experiments on DeepSeek-R1 distilled models validate that our method reduces output tokens by up to 71\\% while maintaining and even improving the accuracy on several mathematical benchmarks. Our method also exhibits robust cross-domain transferability, delivering consistent token reduction performance in code generation and knowledge-based QA tasks. Code is available at: https://github.com/Aries-iai/Manifold_Steering.",
      "url": "https://openreview.net/forum?id=49Rc51iCso",
      "year": 2025,
      "venue": "NIPS 2025",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "49Rc51iCso",
      "track": "main",
      "status": "Poster",
      "keywords": "Large Reasoning Models;Overthinking;Mechanistic Interpretability;Manifold Steering",
      "tldr": "",
      "primary_area": "deep_learning",
      "similarity_score": 11.206476309459948,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.206476309459948,
      "combined_score": 0.0,
      "rank": 59
    },
    {
      "title": "Score-based Pullback Riemannian Geometry: Extracting the Data Manifold Geometry using Anisotropic Flows",
      "authors": [
        "Willem Diepeveen",
        "Georgios Batzolis",
        "Zakhar Shumaylov",
        "Carola-Bibiane Schönlieb"
      ],
      "abstract": "Data-driven Riemannian geometry has emerged as a powerful tool for interpretable representation learning, offering improved efficiency in downstream tasks. Moving forward, it is crucial to balance cheap manifold mappings with efficient training algorithms. In this work, we integrate concepts from pullback Riemannian geometry and generative models to propose a framework for data-driven Riemannian geometry that is scalable in both geometry and learning: score-based pullback Riemannian geometry. Focusing on unimodal distributions as a first step, we propose a score-based Riemannian structure with closed-form geodesics that pass through the data probability density. With this structure, we construct a Riemannian autoencoder (RAE) with error bounds for discovering the correct data manifold dimension. This framework can naturally be used with anisotropic normalizing flows by adopting isometry regularization during training. Through numerical experiments on diverse datasets, including image data, we demonstrate that the proposed framework produces high-quality geodesics passing through the data support, reliably estimates the intrinsic dimension of the data manifold, and provides a global chart of the manifold. To the best of our knowledge, this is the first scalable framework for extracting the complete geometry of the data manifold.",
      "url": "https://icml.cc/virtual/2025/poster/46179",
      "year": 2025,
      "venue": "ICML 2025",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=AJN5btaqNk",
      "citations": null,
      "categories": [],
      "id": "AJN5btaqNk",
      "track": "main",
      "status": "Poster",
      "keywords": "data manifold geometry;score-based pullback Riemannian metric;anisotropic normalizing flows;closed-form geodesics;intrinsic dimension estimation;Riemannian auto-encoder;interpretable representation learning",
      "tldr": "",
      "primary_area": "deep_learning->generative_models_and_autoencoders",
      "similarity_score": 11.153019448311568,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.153019448311568,
      "combined_score": 0.0,
      "rank": 60
    },
    {
      "title": "Rolling Riemannian Manifolds to Solve the Multi-class Classification Problem",
      "authors": [
        "Rui Caseiro",
        "Pedro Martins",
        "Joao F. Henriques",
        "Fatima Silva Leite",
        "Jorge Batista"
      ],
      "abstract": "In the past few years there has been a growing interest on geometric frameworks to learn supervised classification models on Riemannian manifolds [31, 27]. A popular framework, valid over any Riemannian manifold, was proposed in [31] for binary classification. Once moving from binary to multi-class classification this paradigm is not valid anymore, due to the spread of multiple positive classes on the manifold [27]. It is then natural to ask whether the multi-class paradigm could be extended to operate on a large class of Riemannian manifolds. We propose a mathematically well-founded classification paradigm that allows to extend the work in [31] to multi-class models, taking into account the structure of the space. The idea is to project all the data from the manifold onto an affine tangent space at a particular point. To mitigate the distortion induced by local diffeomorphisms, we introduce for the first time in the computer vision community a well-founded mathematical concept, so-called Rolling map [21, 16]. The novelty in this alternate school of thought is that the manifold will be firstly rolled (without slipping or twisting) as a rigid body, then the given data is unwrapped onto the affine tangent space, where the classification is performed.",
      "url": "https://openaccess.thecvf.com/content_cvpr_2013/html/Caseiro_Rolling_Riemannian_Manifolds_2013_CVPR_paper.html",
      "year": 2013,
      "venue": "CVPR 2013",
      "source": "offline_cvpr",
      "doi": null,
      "pdf_url": "https://openaccess.thecvf.com/content_cvpr_2013/papers/Caseiro_Rolling_Riemannian_Manifolds_2013_CVPR_paper.pdf",
      "citations": null,
      "categories": [],
      "id": "89764a0f6a",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 11.115284931622345,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.115284931622345,
      "combined_score": 0.0,
      "rank": 61
    },
    {
      "title": "Parametric Manifold Learning Via Sparse Multidimensional Scaling",
      "authors": [
        "Gautam Pai",
        "Ronen Talmon",
        "Ron Kimmel"
      ],
      "abstract": "We propose a metric-learning framework for computing distance-preserving maps that generate low-dimensional embeddings for a certain class of manifolds. We employ Siamese networks to solve the problem of least squares multidimensional scaling for generating mappings that preserve geodesic distances on the manifold. In contrast to previous parametric manifold learning methods we show a substantial reduction in training effort enabled by the computation of geodesic distances in a farthest point sampling strategy. Additionally, the use of a network to model the distance-preserving map reduces the complexity of the multidimensional scaling problem and leads to an improved non-local generalization of the manifold compared to analogous non-parametric counterparts. We demonstrate our claims on point-cloud data and on image manifolds and show a numerical analysis of our technique to facilitate a greater understanding of the representational power of neural networks in modeling manifold data.",
      "url": "https://openreview.net/forum?id=B1uvH_gC-",
      "year": 2018,
      "venue": "ICLR 2018",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "B1uvH_gC-",
      "track": "main",
      "status": "Reject",
      "keywords": "Manifold Learning;Non-linear Dimensionality Reduction;Neural Networks;Unsupervised Learning",
      "tldr": "Parametric Manifold Learning with Neural Networks in a Geometric Framework ",
      "primary_area": "",
      "similarity_score": 11.096041973713113,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 11.096041973713113,
      "combined_score": 0.0,
      "rank": 62
    },
    {
      "title": "Learning Iterative Neural Optimizers for Image Steganography",
      "authors": [
        "Xiangyu Chen",
        "Varsha Kishore",
        "Kilian Q Weinberger"
      ],
      "abstract": "Image steganography is the process of concealing secret information in images through imperceptible changes. \nRecent work has formulated this task as a classic constrained optimization problem. In this paper, we argue that image steganography is inherently performed on the (elusive) manifold of natural images, and propose an iterative neural network trained to perform the optimization steps. In contrast to classical optimization methods like L-BFGS or projected gradient descent, we train the neural network to also stay close to the manifold of natural images throughout the optimization. We show that our learned neural optimization is faster and more reliable than classical optimization approaches. In comparison to previous state-of-the-art encoder-decoder based steganography methods, it reduces the recovery error rate by multiple orders of magnitude and achieves zero error up to 3 bits per pixel (bpp) without the need for error-correcting codes. ",
      "url": "https://iclr.cc/virtual/2023/poster/10886",
      "year": 2023,
      "venue": "ICLR 2023",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=gLPkzWjdhBN",
      "citations": null,
      "categories": [],
      "id": "gLPkzWjdhBN",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.969882022773383,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.969882022773383,
      "combined_score": 0.0,
      "rank": 63
    },
    {
      "title": "Minimum Curvature Manifold Learning",
      "authors": [
        "Yonghyeon Lee",
        "Frank C. Park"
      ],
      "abstract": "It is widely observed that vanilla autoencoders can have low manifold learning accuracy given a noisy or small training dataset. \nRecent work has discovered that it is important to regularize the decoder that explicitly parameterizes the manifold, \nwhere a neighborhood graph is employed for decoder regularization. However, one caveat of this method is that it is not always straightforward to construct a correct graph. Alternatively, one may consider naive graph-free regularization methods such as minimizing the norm of the decoder's Jacobian or Hessian, but these norms are not coordinate-invariant (i.e. reparametrization-invariant) and hence do not capture any meaningful geometric quantity of the manifold nor result in geometrically meaningful manifold regularization effects. \nAnother recent work called the isometric regularization implicitly forces the manifold to have zero intrinsic curvature, resulting in some geometrically meaningful regularization effects. But, since the intrinsic curvature does not capture how the manifold is embedded in the data space from an extrinsic perspective, the regularization effects are often limited. In this paper, we propose a {\\it minimum extrinsic curvature principle} for manifold regularization and {\\bf Minimum Curvature Autoencoder (MCAE)}, a graph-free coordinate-invariant extrinsic curvature minimization framework for autoencoder regularization. Experiments with various standard datasets demonstrate that MCAE improves manifold learning accuracy compared to existing methods, especially showing strong robustness to noise.",
      "url": "https://openreview.net/forum?id=yxj33c6NuX",
      "year": 2023,
      "venue": "ICLR 2023",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "yxj33c6NuX",
      "track": "main",
      "status": "Reject",
      "keywords": "Autoencoder;Manifold;Curvature;Riemannian geometry",
      "tldr": "We propose a minimum extrinsic curvature principle for manifold regularization and Minimum Curvature Autoencoder (MCAE), a graph-free coordinate-invariant extrinsic curvature minimization framework for autoencoder regularization.",
      "primary_area": "",
      "similarity_score": 10.953067538334954,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.953067538334954,
      "combined_score": 0.0,
      "rank": 64
    },
    {
      "title": "Sparse Representation Classification With Manifold Constraints Transfer",
      "authors": [
        "Baochang Zhang",
        "Alessandro Perina",
        "Vittorio Murino",
        "Alessio Del Bue"
      ],
      "abstract": "The fact that image data samples lie on a manifold has been successfully exploited in many learning and inference problems. In this paper we leverage the specific structure of data in order to improve recognition accuracies in general recognition tasks. In particular we propose a novel framework that allows to embed manifold priors into sparse representation-based classification (SRC) approaches. We also show that manifold constraints can be transferred from the data to the optimized variables if these are linearly correlated. Using this new insight, we define an efficient alternating direction method of multipliers (ADMM) that can consistently integrate the manifold constraints during the optimization process. This is based on the property that we can recast the problem as the projection over the manifold via a linear embedding method based on the Geodesic distance. The proposed approach is successfully applied on face, digit, action and objects recognition showing a consistently increase on performance when compared to the state of the art.",
      "url": "https://openaccess.thecvf.com/content_cvpr_2015/html/Zhang_Sparse_Representation_Classification_2015_CVPR_paper.html",
      "year": 2015,
      "venue": "CVPR 2015",
      "source": "offline_cvpr",
      "doi": null,
      "pdf_url": "https://openaccess.thecvf.com/content_cvpr_2015/papers/Zhang_Sparse_Representation_Classification_2015_CVPR_paper.pdf",
      "citations": null,
      "categories": [],
      "id": "ad8c5bbd1c",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.92814262875056,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.92814262875056,
      "combined_score": 0.0,
      "rank": 65
    },
    {
      "title": "Neural Superposition Networks",
      "authors": [
        "Atiyo Ghosh",
        "Nicolò Toscano",
        "Jongyeong Lee",
        "Hyukgeun Cha",
        "Jun-Ho Lee",
        "Jung Jun Park",
        "Yunjun Choi",
        "Seong-Hyok Sean Kim",
        "Antonio Andrea Gentile"
      ],
      "abstract": "We introduce _Neural Superposition Networks_, a class of physics-constrained neural architectures that exactly satisfy given partial differential equations (PDEs) by construction. In contrast to traditional physics-informed neural networks (PINNs), which enforce PDE constraints via loss regularization, our approach embeds the solution manifold directly into the architecture by expressing the output as a superposition of analytical basis functions that solve the target PDE. This eliminates the need for interior residual loss terms, simplifies training to a single-objective optimization on boundary conditions, and improves numerical stability. \nWe show that for linear PDEs—including Laplace, heat, and incompressible flow constraints—this architectural bias leads to provably convergent approximations. Using maximum principles and classical convergence theory, we establish uniform boundary-to-interior convergence guarantees. For nonlinear PDEs such as Burgers’ equation, we demonstrate that partial structural constraints can still be enforced via transformations (e.g., Cole–Hopf), yielding improved inductive bias over standard PINNs. The resulting networks combine the expressiveness of deep learning with the convergence guarantees of Galerkin and spectral methods. Our framework offers a theoretically grounded and computationally efficient alternative to residual-based training for PDE-constrained problems..",
      "url": "https://openreview.net/forum?id=aWXrVm07Zl",
      "year": 2025,
      "venue": "NIPS 2025",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "aWXrVm07Zl",
      "track": "main",
      "status": "Reject",
      "keywords": "differential equations;physics-informed neural networks;scientific machine learning;differentially constrained architecture;principle of superposition",
      "tldr": "",
      "primary_area": "machine_learning_for_sciences",
      "similarity_score": 10.852639611952323,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.852639611952323,
      "combined_score": 0.0,
      "rank": 66
    },
    {
      "title": "Universality Theorems for Generative Models",
      "authors": [
        "Valentin Khrulkov",
        "Ivan Oseledets"
      ],
      "abstract": "Despite the fact that generative models are extremely successful in practice, the theory underlying this phenomenon is only starting to catch up with practice. In this work we address the question of the universality of generative models: is it true that neural networks can approximate any data manifold arbitrarily well? We provide a positive answer to this question and show that under mild assumptions on the activation function one can always find a feedforward neural network that maps the latent space onto a set located within the specified Hausdorff distance from the desired data manifold. We also prove similar theorems for the case of multiclass generative models and cycle generative models, trained to map samples from one manifold to another and vice versa.",
      "url": "https://openreview.net/forum?id=rJlJF1SYPB",
      "year": 2020,
      "venue": "ICLR 2020",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "rJlJF1SYPB",
      "track": "main",
      "status": "Withdraw",
      "keywords": "generative models;theory;universality;manifolds;differential geometry",
      "tldr": "We shot that a wide class of manifolds can be generated by ReLU and sigmoid networks with arbitrary precision.",
      "primary_area": "",
      "similarity_score": 10.844474991951449,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.844474991951449,
      "combined_score": 0.0,
      "rank": 67
    },
    {
      "title": "An Inexact Regularized Adaptive Algorithm with Manifold Identification for Training Structured Neural Networks",
      "authors": [
        "Zih-Syuan Huang",
        "Ching-pei Lee"
      ],
      "abstract": "We propose an inexact regularized adaptive dual averaging algorithm with momentum, RAMDA, for training structured neural networks in various tasks with the help of regularization. Through the theory of manifold identification, we show that after a finite number of steps, the structures of the iterates generated by RAMDA are all identical to the structure induced by the regularization at the stationary point of asymptotic convergence.  This structure is locally optimal within a neighborhood of the point of convergence and therefore provides the best possible performance among all methods converging to the same point. To make use of manifold identification, RAMDA produces stochastic estimators of the gradient that almost surely converge to the true gradient even when the training problem is no longer a finite-sum one but a stochastic one over a certain probability distribution due to data augmentation. With the simultaneous presence of a preconditioner and a regularization term, the subproblem of RAMDA as well as those of existing frameworks have no closed-form solutions, so we also propose a general iterative subroutine for approximately solving such subproblems efficiently while maintaining similar convergence guarantees. Extensive numerical experiments in modern computer vision, natural language processing, and speech tasks show that our subproblem solver is efficient and applicable to existing frameworks, and the proposed RAMDA excels state of the art for training structured neural networks to generate more structural points without decreasing the prediction performance.",
      "url": "https://openreview.net/forum?id=BlCnycxgJQ",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "BlCnycxgJQ",
      "track": "main",
      "status": "Reject",
      "keywords": "Deep learning;structured models;adaptive method;manifold identification;variance reduction;inexact subproblem solution",
      "tldr": "",
      "primary_area": "optimization",
      "similarity_score": 10.832179887472531,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.832179887472531,
      "combined_score": 0.0,
      "rank": 68
    },
    {
      "title": "REMA: A Unified Reasoning Manifold Framework for Interpreting Large Language Model",
      "authors": [],
      "abstract": "Understanding how Large Language Models (LLMs) perform complex reasoning and their failure mechanisms is a challenge in interpretability research.\nTo provide a measurable geometric analysis perspective, we define the concept of the **Reasoning Manifold**, a latent low-dimensional geometric structure formed by the internal representations corresponding to all correctly reasoned generations. \nThis structure can be conceptualized as the embodiment of the effective thinking paths that the model has learned to successfully solve a given task.\nBased on this concept, we build **REMA**, a framework that explains the origins of failures by quantitatively comparing the spatial relationships of internal model representations corresponding to both erroneous and correct reasoning samples.\nSpecifically, REMA first quantifies the geometric deviation of each erroneous representation by calculating its k-nearest neighbors distance to the approximated manifold formed by correct representations, thereby providing a unified failure signal.\nIt then localizes the divergence points where these deviations first become significant by tracking this deviation metric across the model's layers and comparing it against a baseline of internal fluctuations from correct representations, thus identifying where the reasoning chain begins to go off-track.\nOur extensive experiments on diverse language and multimodal models and tasks demonstrate the low-dimensional nature of the reasoning manifold and the high separability between erroneous and correct reasoning representations.\nThe results also validate the effectiveness of the REMA framework in analyzing the origins of reasoning failures.\nThis research connects abstract reasoning failures to measurable geometric deviations in representations, providing new avenues for in-depth understanding and diagnosis of the internal computational processes of black-box models.",
      "url": "https://openreview.net/forum?id=A8ez8ThZWq",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "A8ez8ThZWq",
      "track": "main",
      "status": "Active",
      "keywords": "Interpretability;Large Language Models;Reasoning Manifold",
      "tldr": "",
      "primary_area": "interpretability and explainable AI",
      "similarity_score": 10.820747705358734,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.820747705358734,
      "combined_score": 0.0,
      "rank": 69
    },
    {
      "title": "Matrix Tri-Factorization With Manifold Regularizations for Zero-Shot Learning",
      "authors": [
        "Xing Xu",
        "Fumin Shen",
        "Yang Yang",
        "Dongxiang Zhang",
        "Heng Tao Shen",
        "Jingkuan Song"
      ],
      "abstract": "Zero-shot learning (ZSL) aims to recognize objects of unseen classes with available training data from another set of seen classes. Existing solutions are focused on exploring knowledge transfer via an intermediate semantic embedding (e.g.s, attributes) shared between seen and unseen classes. In this paper, we propose a novel projection framework based on matrix tri-factorization with manifold regularizations. Specifically, we learn the semantic embedding projection by decomposing the visual feature matrix under the guidance of semantic embedding and class label matrices. By additionally introducing manifold regularizations on visual data and semantic embeddings, the learned projection can effectively captures the geometrical manifold structure residing in both visual and semantic spaces. To avoid the projection domain shift problem, we devise an effective prediction scheme by exploiting the test-time manifold structure. Extensive experiments on four benchmark datasets show that our approach significantly outperforms the state-of-the-arts, yielding an average improvement  ratio by 7.4% and 31.9% for the recognition and retrieval task, respectively.",
      "url": "",
      "year": 2017,
      "venue": "CVPR 2017",
      "source": "offline_cvpr",
      "doi": null,
      "pdf_url": "https://openaccess.thecvf.com/content_cvpr_2017/papers/Xu_Matrix_Tri-Factorization_With_CVPR_2017_paper.pdf",
      "citations": null,
      "categories": [],
      "id": "",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.813481201742702,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.813481201742702,
      "combined_score": 0.0,
      "rank": 70
    },
    {
      "title": "Zero Shot Learning via Multi-Scale Manifold Regularization",
      "authors": [
        "Shay Deutsch",
        "Soheil Kolouri",
        "Kyungnam Kim",
        "Yuri Owechko",
        "Stefano Soatto"
      ],
      "abstract": "We address zero-shot learning using a new manifold alignment framework based on a localized multi-scale transform on graphs. Our inference approach includes a smoothness criterion for a function mapping nodes on a graph (visual representation) onto a linear space (semantic representation), which we optimize using multi-scale graph wavelets. The robustness of the ensuing scheme allows us to operate with automatically generated semantic annotations, resulting in an algorithm that is entirely free of manual supervision, and yet improves the state-of-the-art as measured on benchmark datasets.",
      "url": "",
      "year": 2017,
      "venue": "CVPR 2017",
      "source": "offline_cvpr",
      "doi": null,
      "pdf_url": "https://openaccess.thecvf.com/content_cvpr_2017/papers/Deutsch_Zero_Shot_Learning_CVPR_2017_paper.pdf",
      "citations": null,
      "categories": [],
      "id": "",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.770908213950886,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.770908213950886,
      "combined_score": 0.0,
      "rank": 71
    },
    {
      "title": "Differential Privacy with Manifold Data Dependency",
      "authors": [
        "Lei Wang",
        "Deming Yuan",
        "Guodong Shi"
      ],
      "abstract": "In this paper, we study dataset processing mechanisms generated by linear queries in the presence of manifold data dependency. Specifically, the input data are assumed to lie in an affine manifold as prior knowledge known to adversaries. First of all, we show such manifold data dependency may have a significant impact on the privacy levels compared to the case with the manifold constraint being absent.  We establish necessary and sufficient conditions on the possibility of achieving differential privacy  via structured noise injection mechanisms where  non i.i.d. Gaussian or Laplace noises are calibrated into dataset. Next, in light of these conditions, procedures are developed  by which a prescribed privacy budget  can be tightly reached with a matching noise level.  Finally, we show that the  framework has immediate applications  in  differentially private cloud-based control, where the manifold data dependency arises naturally from the system dynamics, and the proposed theories and procedures become effective tools in evaluating privacy levels and in the design of provably useful  algorithms.",
      "url": "https://openreview.net/forum?id=zokEN0xOb0Q",
      "year": 2022,
      "venue": "ICLR 2022",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "zokEN0xOb0Q",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Differential privacy;data correlation",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.763307789258212,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.763307789258212,
      "combined_score": 0.0,
      "rank": 72
    },
    {
      "title": "Neural Collapse with Normalized Features: A Geometric Analysis over the Riemannian Manifold",
      "authors": [
        "Can Yaras",
        "Peng Wang",
        "Zhihui Zhu",
        "Laura Balzano",
        "Qing Qu"
      ],
      "abstract": "When training overparameterized deep networks for classification tasks, it has been widely observed that the learned features exhibit a so-called \"neural collapse'\" phenomenon. More specifically, for the output features of the penultimate layer, for each class the within-class features converge to their means, and the means of different classes exhibit a certain tight frame structure, which is also aligned with the last layer's classifier. As feature normalization in the last layer becomes a common practice in modern representation learning, in this work we theoretically justify the neural collapse phenomenon under normalized features. Based on an unconstrained feature model, we simplify the empirical loss function in a multi-class classification task into a nonconvex optimization problem over the Riemannian manifold by constraining all features and classifiers over the sphere. In this context, we analyze the nonconvex landscape of the Riemannian optimization problem over the product of spheres, showing a benign global landscape in the sense that the only global minimizers are the neural collapse solutions while all other critical points are strict saddle points with negative curvature. Experimental results on practical deep networks corroborate our theory and demonstrate that better representations can be learned faster via feature normalization. Code for our experiments can be found at https://github.com/cjyaras/normalized-neural-collapse.",
      "url": "https://nips.cc/virtual/2022/poster/54456",
      "year": 2022,
      "venue": "NIPS 2022",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=Zvh6lF5b26N",
      "citations": null,
      "categories": [],
      "id": "Zvh6lF5b26N",
      "track": "main",
      "status": "Accept",
      "keywords": "neural collapse;Riemannian manifold;feature normalization;nonconvex optimization",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.759296987419702,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.759296987419702,
      "combined_score": 0.0,
      "rank": 73
    },
    {
      "title": "Log-Hilbert-Schmidt metric between positive definite operators on Hilbert spaces",
      "authors": [
        "Hà Quang Minh",
        "Marco San Biagio",
        "Vittorio Murino"
      ],
      "abstract": "This paper introduces a novel mathematical and computational framework, namely {\\it Log-Hilbert-Schmidt metric} between positive definite operators on a Hilbert space. This is a generalization of the Log-Euclidean metric on the Riemannian manifold of positive definite matrices to the infinite-dimensional setting. The general framework is applied in particular to compute distances between covariance operators on a Reproducing Kernel Hilbert Space (RKHS), for which we obtain explicit formulas via the corresponding Gram matrices. Empirically, we apply our formulation to the task of multi-category image classification, where each image is represented by an infinite-dimensional RKHS covariance operator. On several challenging datasets, our method significantly outperforms approaches based on covariance matrices computed directly on the original input features, including those using the Log-Euclidean metric, Stein and Jeffreys divergences, achieving new state of the art results.",
      "url": "https://nips.cc/virtual/2014/poster/4408",
      "year": 2014,
      "venue": "NIPS 2014",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2014/file/3000e56b48442cd23b49e5064bf1a9e6-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "4408",
      "track": "main",
      "status": "Spotlight",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.753793156063207,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.753793156063207,
      "combined_score": 0.0,
      "rank": 74
    },
    {
      "title": "Graph-Enhanced Learning for Predicting Optimal Drug Combinations Using Contrastive Embedding",
      "authors": [
        "Zhenghan chen",
        "youhuan yang",
        "Lang Zheng",
        "Ruxue Xing",
        "Han Quan"
      ],
      "abstract": "We present a groundbreaking unified theory for drug-drug interaction (DDI) aware domain adaptation (DA) in the context of drug synergy prediction. Our framework seamlessly integrates concepts from optimal transport, information geometry, and quantum information theory within the setting of abstract Banach spaces. We introduce a novel DDI-aware optimal transport problem, formulated as a geodesic equation on an infinite-dimensional Finsler manifold that encodes both DDI structure and optimal transport costs. This geometric formulation provides a unified perspective on DDI-aware domain adaptation, interpreting the process as the evolution of a transport map along a geodesic in a space that captures both domain discrepancy and drug interaction patterns. Our approach extends to a stochastic gradient flow on the space of probability measures, combining ideas from information geometry and stochastic analysis. We prove the existence of a unique invariant measure for this flow and establish its convergence properties using techniques from infinite-dimensional Markov processes and Γ-convergence. Our comprehensive mathematical framework not only unifies existing approaches to domain adaptation and DDI prediction but also opens new avenues for research at the intersection of these fields. By bridging the gap between abstract mathematical theories and practical drug synergy prediction, our work paves the way for more effective and theoretically grounded algorithms in drug discovery and personalized medicine. The proposed unified theory has far-reaching implications, potentially revolutionizing our understanding of cross-domain adaptation in complex biochemical systems and inspiring novel computational methods in pharmaceutical research.",
      "url": "https://openreview.net/forum?id=plAiJUFNja",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "plAiJUFNja",
      "track": "main",
      "status": "Reject",
      "keywords": "Graph Learning;Contrastive Embedding;DDI",
      "tldr": "",
      "primary_area": "transfer learning, meta learning, and lifelong learning",
      "similarity_score": 10.74504625122039,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.74504625122039,
      "combined_score": 0.0,
      "rank": 75
    },
    {
      "title": "Generative Counterfactual Manifold Perturbation: A Robust Framework for Treatment Effect Estimation with Unobserved Confounders",
      "authors": [],
      "abstract": "Estimating treatment effects from observational data is difficult when unobserved confounders create spurious associations that bias simple estimators. Recent generative approaches learn outcome distributions with conditional diffusion models, and some robust representation methods introduce sensitivity analysis or structural priors. These advances work well when identification assumptions hold exactly, but they become fragile when those assumptions are only approximate and offer few practical diagnostics. We introduce Generative Counterfactual Manifold Perturbation (GCMP), a unified framework that combines causal-aware self supervised learning, conditional diffusion counterfactual proxy generation, and adaptive variational inference. GCMP makes three main contributions: (i) a self supervised objective that preserves confounding signals during representation learning; (ii) a conditional diffusion model that reframes proxy construction as a generative task over rich perturbation manifolds; and (iii) an adaptive regularization scheme that yields graceful degradation and calibrated uncertainty when identification assumptions are violated. We also present new identifiability conditions, finite sample error bounds, and diagnostic tests to quantify manifold quality and effective orthogonality. Extensive experiments on synthetic and semi-synthetic benchmarks show that GCMP consistently outperforms the state-of-the-art.",
      "url": "https://openreview.net/forum?id=JZ3Svjj9hG",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "JZ3Svjj9hG",
      "track": "main",
      "status": "Active",
      "keywords": "ML: Causal Learning",
      "tldr": "",
      "primary_area": "causal reasoning",
      "similarity_score": 10.744096254890696,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.744096254890696,
      "combined_score": 0.0,
      "rank": 76
    },
    {
      "title": "Learning Manifolds with K-Means and K-Flats",
      "authors": [
        "Guillermo Canas",
        "Tomaso Poggio",
        "Lorenzo Rosasco"
      ],
      "abstract": "We study the problem of estimating a manifold from random samples. In particular, we consider piecewise constant and piecewise linear estimators induced by  k-means and k-flats, and analyze their performance. We extend previous results  for k-means in two separate directions. First, we provide new results for k-means reconstruction on manifolds and, secondly, we prove reconstruction bounds for higher-order approximation (k-flats), for which no known results were previously available. While the results for k-means are novel, some of the technical tools are well-established in the literature. In the case of k-flats, both the results and the mathematical tools are  new.",
      "url": "https://papers.nips.cc/paper_files/paper/2012/hash/b20bb95ab626d93fd976af958fbc61ba-Abstract.html",
      "year": 2012,
      "venue": "NIPS 2012",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2012/file/b20bb95ab626d93fd976af958fbc61ba-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "f15e9021c5",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.721056406139137,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.721056406139137,
      "combined_score": 0.0,
      "rank": 77
    },
    {
      "title": "Unified K-Means Clustering with Label-Guided Manifold Learning",
      "authors": [
        "Qianqian Wang",
        "Mengping Jiang",
        "Zhengming Ding",
        "Quanxue Gao"
      ],
      "abstract": "K-Means clustering is a classical and effective unsupervised learning method attributed to its simplicity and efficiency. However, it faces notable challenges, including sensitivity to random initial centroid selection, a limited ability to discover the intrinsic manifold structures within nonlinear datasets, and difficulty in achieving balanced clustering in practical scenarios. To overcome these weaknesses, we introduce a novel framework for K-Means that leverages manifold learning. This approach eliminates the need for centroid calculation and utilizes a cluster indicator matrix to align the manifold structures, thereby enhancing clustering accuracy. Beyond the traditional Euclidean distance, our model incorporates Gaussian kernel distance, K-nearest neighbor distance, and low-pass filtering distance to effectively manage data that is not linearly separable. Furthermore, we introduce a balanced regularizer to achieve balanced clustering results. The detailed experimental results demonstrate the efficacy of our proposed methodology.",
      "url": "https://icml.cc/virtual/2025/poster/44925",
      "year": 2025,
      "venue": "ICML 2025",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=YD9ZoqUDAY",
      "citations": null,
      "categories": [],
      "id": "YD9ZoqUDAY",
      "track": "main",
      "status": "Poster",
      "keywords": "Balanced clustering;unsupervised learning;low-pass filtering distance.",
      "tldr": "",
      "primary_area": "general_machine_learning->clustering",
      "similarity_score": 10.704778930496667,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.704778930496667,
      "combined_score": 0.0,
      "rank": 78
    },
    {
      "title": "Learning Manifold Implicitly via Explicit Heat-Kernel Learning",
      "authors": [
        "Yufan Zhou",
        "Changyou Chen",
        "Jinhui Xu"
      ],
      "abstract": "Manifold learning is a fundamental problem in machine learning with numerous applications. Most of the existing methods directly learn the low-dimensional embedding of the data in some high-dimensional space, and usually lack the flexibility of being directly applicable to down-stream applications. In this paper, we propose the concept of implicit manifold learning, where manifold information is implicitly obtained by learning the associated heat kernel. A heat kernel is the solution of the corresponding heat equation, which describes how ``heat'' transfers on the manifold, thus containing ample geometric information of the manifold. We provide both practical algorithm and theoretical analysis of our framework. The learned heat kernel can be applied to various kernel-based machine learning models, including deep generative models (DGM) for data generation and Stein Variational Gradient Descent for Bayesian inference. Extensive experiments show that our framework can achieve the state-of-the-art results compared to existing methods for the two tasks.",
      "url": "https://nips.cc/virtual/2020/poster/18491",
      "year": 2020,
      "venue": "NIPS 2020",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2020/file/05e2a0647e260c355dd2b2175edb45b8-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "18491",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.697123240237127,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.697123240237127,
      "combined_score": 0.0,
      "rank": 79
    },
    {
      "title": "Varying Manifolds in Diffusion: From Time-varying Geometries to Visual Saliency",
      "authors": [
        "Junhao Chen",
        "Manyi Li",
        "zherong pan",
        "Xifeng Gao",
        "Changhe Tu"
      ],
      "abstract": "Building on the manifold hypothesis, which suggests that generative models learn data distributions residing on low-dimensional manifolds, this paper investigates the time-varying manifold sequence induced by the generation process through the lens of differential equations in diffusion models. Our primary contribution is the introduction of the \\textit{generation rate}, a novel metric that quantifies local manifold scaling over time. For image data, we show that the accumulated generation rate, referred to as the \\textit{generation curve}, strongly correlates with intuitive visual properties, such as the saliency of image components. By leveraging modifications to the generation curves, we propose a unified framework for a range of image manipulation tasks, including semantic transfer, object removal, saliency adjustment, and image blending. Comprehensive evaluations, supported by both the qualitative and quantitative results, highlight the effectiveness of our framework across these diverse tasks.",
      "url": "https://openreview.net/forum?id=mIGCz3ZmmX",
      "year": 2025,
      "venue": "ICML 2025",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "mIGCz3ZmmX",
      "track": "main",
      "status": "Reject",
      "keywords": "Manifold analysis; Diffusion model; Image manipuation",
      "tldr": "",
      "primary_area": "deep_learning->generative_models_and_autoencoders",
      "similarity_score": 10.582027579717925,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.582027579717925,
      "combined_score": 0.0,
      "rank": 80
    },
    {
      "title": "A TWO-STAGE FRAMEWORK FOR MATHEMATICAL EXPRESSION RECOGNITION",
      "authors": [
        "Jin Zhang",
        "Weipeng Ming",
        "Pengfei Liu"
      ],
      "abstract": "\nAlthough mathematical expressions (MEs) recognition have achieved great progress, the development of MEs recognition in real scenes is still unsatisfactory. Inspired by the recent work of neutral network, this paper proposes a novel two-stage approach which takes a printed mathematical expression image as input and generates LaTeX sequence as output. In the first stage, this method locates and recognizes the math symbols of input image by object detection algorithm. In the second stage, it translates math symbols with position information into LaTeX sequences by seq2seq model equipped with attention mechanism. In particular, the detection of mathematical symbols and the structural analysis of mathematical formulas are carried out separately in two steps, which effectively improves the recognition accuracy and enhances the generalization ability. The experiment demonstrates that the two-stage method significantly outperforms the end-to-end method. Especially, the ExpRate(expression recognition rate) of our model is 74.1%, 20.3 percentage points higher than that of the end-to-end model on the test data that doesn’t come from the same source as training data.",
      "url": "https://openreview.net/forum?id=rkl_Ch4YwS",
      "year": 2020,
      "venue": "ICLR 2020",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "rkl_Ch4YwS",
      "track": "main",
      "status": "Reject",
      "keywords": "mathematical expressions recognition;seq2seq model",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.577119675785104,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.577119675785104,
      "combined_score": 0.0,
      "rank": 81
    },
    {
      "title": "WarriorMath: Empowering Mathematical Reasoning for Large Language Models via Expert Battles",
      "authors": [],
      "abstract": "Large Language Models (LLMs) excel in solving mathematical problems, yet their performance is often limited by the availability of high-quality, diverse training data. Existing methods focus on augmenting datasets through rephrasing or difficulty progression but overlook the specific failure modes of LLMs. This results in synthetic questions that the model can already solve, providing minimal performance gains. To address this, we propose WarriorMath, a defect-aware framework for mathematical problem solving that integrates both targeted data synthesis and progressive training. In the synthesis stage, we employ multiple expert LLMs in a collaborative process to generate, critique, and refine problems. Questions that base LLMs fail to solve are identified and iteratively improved through expert-level feedback, producing high-quality, defect-aware training data. In the training stage, we introduce a progressive learning framework that iteratively fine-tunes the model using increasingly challenging data tailored to its weaknesses. Experiments on six mathematical benchmarks show that WarriorMath outperforms strong baselines by 12.57% on average, setting a new state-of-the-art. Our results demonstrate the effectiveness of a defect-aware, multi-expert framework for improving mathematical ability.",
      "url": "https://openreview.net/forum?id=JPZoLWYo82",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "JPZoLWYo82",
      "track": "main",
      "status": "Active",
      "keywords": "Mathematical NLP;Data Synthesis",
      "tldr": "",
      "primary_area": "foundation or frontier models, including LLMs",
      "similarity_score": 10.566467573485692,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.566467573485692,
      "combined_score": 0.0,
      "rank": 82
    },
    {
      "title": "Modelling the influence of data structure on learning in neural networks",
      "authors": [
        "S. Goldt",
        "M. Mézard",
        "F. Krzakala",
        "L. Zdeborová"
      ],
      "abstract": "The lack of crisp mathematical models that capture the structure of real-world\ndata sets is a major obstacle to the detailed theoretical understanding of deep\nneural networks. Here, we first demonstrate the effect of structured data sets\nby experimentally comparing the dynamics and the performance of two-layer\nnetworks trained on two different data sets: (i) an unstructured synthetic data\nset containing random i.i.d. inputs, and (ii) a simple canonical data set such\nas MNIST images. Our analysis reveals two phenomena related to the dynamics of\nthe networks and their ability to generalise that only appear when training on\nstructured data sets. Second, we introduce a generative model for data sets,\nwhere high-dimensional inputs lie on a lower-dimensional manifold and have\nlabels that depend only on their position within this manifold. We call it the\n*hidden manifold model* and we experimentally demonstrate that training\nnetworks on data sets drawn from this model reproduces both the phenomena seen\nduring training on MNIST.",
      "url": "https://openreview.net/forum?id=BJlisySYPS",
      "year": 2020,
      "venue": "ICLR 2020",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "BJlisySYPS",
      "track": "main",
      "status": "Reject",
      "keywords": "Neural Networks;Generative models;Synthetic data sets;Generalisation;Stochastic Gradient descent",
      "tldr": "We demonstrate how structure in data sets impacts neural networks and introduce a generative model for synthetic data sets that reproduces this impact.",
      "primary_area": "",
      "similarity_score": 10.555737883137567,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.555737883137567,
      "combined_score": 0.0,
      "rank": 83
    },
    {
      "title": "MedCalc-R1: Knowledge-Guided Reward Framework for Medical Mathematical Reasoning",
      "authors": [],
      "abstract": "Medical mathematical reasoning is a critical component of clinical decision-making, where accuracy directly affects patient safety and treatment outcomes. However, existing large model approaches, while improving complex reasoning ability, often suffer from knowledge degradation, computational bias, and lack of interpretability. Moreover, commonly used reward mechanisms rely heavily on coarse-grained acceptable ranges, which fail to guarantee stable and precise mathematical outputs. To address these challenges, we propose a knowledge-guided reward framework with two complementary mechanisms. First, a knowledge verification reward enforces explicit formula generation and leverages an independent verification model to check both formulas and results, thereby mitigating knowledge forgetting, enhancing interpretability, and improving reasoning transparency. Second, a hybrid soft–hard reward mechanism incorporates clinical safety thresholds as hard constraints and introduces progressive accuracy-based rewards as soft optimization, simultaneously achieving improvements in both safety and precision. Extensive experiments on medical mathematical reasoning tasks demonstrate that our approach significantly outperforms existing methods in terms of reasoning accuracy, knowledge robustness, and model generalization, thereby validating the effectiveness and broad applicability of the proposed framework.",
      "url": "https://openreview.net/forum?id=kKvEleeIsa",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "kKvEleeIsa",
      "track": "main",
      "status": "Active",
      "keywords": "medical mathematical reasoning;knowledge-guided reward;complex reasoning;large language model",
      "tldr": "",
      "primary_area": "applications to computer vision, audio, language, and other modalities",
      "similarity_score": 10.527584291256415,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.527584291256415,
      "combined_score": 0.0,
      "rank": 84
    },
    {
      "title": "Manifold denoising by Nonlinear Robust Principal Component Analysis",
      "authors": [
        "He Lyu",
        "Ningyu Sha",
        "Shuyang Qin",
        "Ming Yan",
        "Yuying Xie",
        "Rongrong Wang"
      ],
      "abstract": "This paper extends robust principal component analysis (RPCA) to nonlinear manifolds. Suppose that the observed data matrix is the sum of a sparse component and a component drawn from some low dimensional manifold. Is it possible to separate them by using similar ideas as RPCA? Is there any benefit in treating the manifold as a whole as opposed to treating each local region independently? We answer these two questions affirmatively by proposing and analyzing an optimization framework that separates the sparse component from the manifold under noisy data. Theoretical error bounds are provided when the tangent spaces of the manifold satisfy certain incoherence conditions. We also provide a near optimal choice of the tuning parameters for the proposed optimization formulation with the help of a new curvature estimation method. The efficacy of our method is demonstrated on both synthetic and real datasets.",
      "url": "https://nips.cc/virtual/2019/poster/14278",
      "year": 2019,
      "venue": "NIPS 2019",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2019/file/a76c0abe2b7b1b79e70f0073f43c3b44-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "14278",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.524920802136846,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.524920802136846,
      "combined_score": 0.0,
      "rank": 85
    },
    {
      "title": "Back to the Continuous Attractor",
      "authors": [
        "Ábel Ságodi",
        "Guillermo Martín-Sánchez",
        "Piotr A Sokol",
        "Il Memming Park"
      ],
      "abstract": "Continuous attractors offer a unique class of solutions for storing continuous-valued variables in recurrent system states for indefinitely long time intervals.\nUnfortunately, continuous attractors suffer from severe structural instability in general---they are destroyed by most infinitesimal changes of the dynamical law that defines them.\nThis fragility limits their utility especially in biological systems as their recurrent dynamics are subject to constant perturbations.\nWe observe that the bifurcations from continuous attractors in theoretical neuroscience models display various structurally stable forms.\nAlthough their asymptotic behaviors to maintain memory are categorically distinct, their finite-time behaviors are similar.\nWe build on the persistent manifold theory to explain the commonalities between bifurcations from and approximations of continuous attractors.\nFast-slow decomposition analysis uncovers the existence of a persistent slow manifold that survives the seemingly destructive bifurcation, relating the flow within the manifold to the size of the perturbation. Moreover, this allows the bounding of the memory error of these approximations of continuous attractors.\nFinally, we train recurrent neural networks on analog memory tasks to support the appearance of these systems as solutions and their generalization capabilities.\nTherefore, we conclude that continuous attractors are functionally robust and remain useful as a universal analogy for understanding analog memory.",
      "url": "https://neurips.cc/virtual/2024/poster/94178",
      "year": 2024,
      "venue": "NIPS 2024",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=fvG6ZHrH0B",
      "citations": null,
      "categories": [],
      "id": "fvG6ZHrH0B",
      "track": "main",
      "status": "Poster",
      "keywords": "continuous attractors;robustness;fast-slow decomposition;generalization",
      "tldr": "",
      "primary_area": "neuroscience_and_cognitive_science",
      "similarity_score": 10.481455708806351,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.481455708806351,
      "combined_score": 0.0,
      "rank": 86
    },
    {
      "title": "Learning the energy relaxation manifold from unrelaxed structures with RelaxNet",
      "authors": [],
      "abstract": "In an effort to bypass computationally expensive density functional theory (DFT) calculations for energy minimization and structure relaxation, rapid progress in the development of machine learning force fields/interatomic potentials (MLFF/MLIPs) and more robust models that adhere to quantum chemistry/physical paradigms and constraints have been realized. However, most research to date involves static-frame energy predictions only (i.e., given a specific atomic configuration, predict the energy of the current or final instance), neglecting intermediary physical insight-providing contexts. In this work, we developed RelaxNet, the first end-to-end, dynamics-aware, equivariant model that leverages neural ordinary differential equations (ODEs) and message passing neural networks (MPNNs) for predicting the energy relaxation landscape between the initial unrelaxed structure and final relaxed structure. From just the initial structure, which is often the configuration that is fed into DFT simulations, we can accurately recover the energy, forces, and geometric pathways for the trajectory at a competitive prediction accuracy, as evidenced by comprehensive benchmarking with state-of-the-art static models and MLIP-based relaxation methods. Additionally, we provide extensive insights on the use of implicit vs. explicit latent embedding evolution to offer perspectives on optimal strategies for future works that seek to integrate expensive graph-based neural networks and neural ODEs.",
      "url": "https://openreview.net/forum?id=2NZxmGjDZj",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "2NZxmGjDZj",
      "track": "main",
      "status": "Active",
      "keywords": "neural ODEs;energy minimization;trajectory;relaxation;forcefield;optimization",
      "tldr": "",
      "primary_area": "applications to physical sciences (physics, chemistry, biology, etc.)",
      "similarity_score": 10.47770259720595,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.47770259720595,
      "combined_score": 0.0,
      "rank": 87
    },
    {
      "title": "Reservoir Computing with Spatial Filtering and Manifold Learning for fMRI Classification",
      "authors": [],
      "abstract": "We introduce a parametric framework that couples discriminative spatial filtering with  reservoir computing to distinguish spatiotemporal structure in resting-state fMRI in two classes. Temporal dependencies are encoded in a reservoir, while supervised spatial filtering on reservoir states isolates condition-specific patterns; parametric Uniform Manifold Approximation and Projection (UMAP) then yields compact nonlinear embeddings fit on training data and evaluated with cross-subject validation. On 163 participants (97 healthy controls, 66 major depressive disorder), the method reaches 87\\% accuracy, outperforming network-feature pipelines using LDA, SVM, kNN, and GNN. Interpretability combines spatial-pattern maps with Shapley-value attribution, providing coherent, region-level explanations that consistently implicate cortical and subcortical areas associated with major depressive disorder. The framework offers an interpretable route to modeling spatiotemporal organization in clinical and cognitive fMRI.",
      "url": "https://openreview.net/forum?id=e4KSeTmjAe",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "e4KSeTmjAe",
      "track": "main",
      "status": "Active",
      "keywords": "Reservoir Computing;Common Spatial Patterns;UMAP;fMRI;Classification;Interpretability",
      "tldr": "",
      "primary_area": "applications to neuroscience & cognitive science",
      "similarity_score": 10.475077743191717,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.475077743191717,
      "combined_score": 0.0,
      "rank": 88
    },
    {
      "title": "Fast, Accurate Manifold Denoising by Tunneling Riemannian Optimization",
      "authors": [
        "Shiyu Wang",
        "Mariam Avagyan",
        "Yihan Shen",
        "Arnaud Lamy",
        "Tingran Wang",
        "Szabolcs Marka",
        "Zsuzsanna Marka",
        "John Wright"
      ],
      "abstract": "Learned denoisers play a fundamental role in various signal generation (e.g., diffusion models) and reconstruction (e.g., compressed sensing) architectures, whose success derives from their ability to leverage low-dimensional structure in data. Existing denoising methods, however, either rely on local approximations that require a linear scan of the entire dataset or treat denoising as generic function approximation problems, sacrificing efficiency and interpretability. We consider the problem of efficiently denoising a new noisy data point sampled from an unknown manifold $\\mathcal M \\in \\mathbb{R}^D$, using only noisy samples. This work proposes a framework for test-time efficient manifold denoising, by framing the concept of  \"learning-to-denoise\" as  *\"learning-to-optimize\"*. We have two technical innovations: (i) *online learning* methods which learn to optimize over the manifold of clean signals using only noisy data, effectively \"growing\" an optimizer one sample at a time. (ii) *mixed-order* methods which guarantee that the learned optimizers achieve global optimality, ensuring both efficiency and near-optimal denoising performance. We corroborate these claims with theoretical analyses of both the complexity and denoising performance of mixed-order traversal. Our experiments on scientific manifolds demonstrate significantly improved complexity-performance tradeoffs compared to nearest neighbor search, which underpins existing provable denoising approaches based on exhaustive search.",
      "url": "https://icml.cc/virtual/2025/poster/44296",
      "year": 2025,
      "venue": "ICML 2025",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=jbafwTkVUn",
      "citations": null,
      "categories": [],
      "id": "jbafwTkVUn",
      "track": "main",
      "status": "Poster",
      "keywords": "Manifold Denoising;Learning-to-optimize",
      "tldr": "",
      "primary_area": "optimization->everything_else",
      "similarity_score": 10.467705477368572,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.467705477368572,
      "combined_score": 0.0,
      "rank": 89
    },
    {
      "title": "The data manifold under the microscope",
      "authors": [],
      "abstract": "A significant gap exists between theory and practice in deep learning. One example is given by generalization and approximation error bounds, which are often derived for overly simplified models or yield guarantees that are too loose to be informative. Many such bounds rely on the manifold hypothesis and depend on geometric regularity properties, including intrinsic dimension, curvature, and reach of the data manifold or target functions. To make progress on improving these bounds, one needs detailed insight into data manifold geometry and suitable benchmarks on simple datasets. However, existing datasets and analysis tools typically fall into two extremes: analytically defined manifolds with precisely known geometry but limited realism, or real-world datasets where bounds are assessed only through downstream performance and geometric properties can be estimated only coarsely and with hard-to-quantify error.\n\nTo address this lack of simple yet realistic datasets and accompanying geometric tools, we introduce a benchmarking framework for studying data geometry. We repurpose and extend the dSprites and COIL-20 datasets with additional transformation dimensions and finer sampling resolution. This enables accurate finite-difference estimates of geometric quantities such as curvature, reach, and volume, yielding a flexible benchmark for evaluating manifold learning methods. As illustrative applications, we assess two established manifold learning bounds by Genovese et al. and Fefferman et al., and analyze how manifold geometry evolves across network layers in $\\beta$-VAEs. Our results highlight both the limitations of existing bounds and the value of controlled benchmarks for guiding future theoretical developments.",
      "url": "https://openreview.net/forum?id=Kk08XcQCl2",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "Kk08XcQCl2",
      "track": "main",
      "status": "Active",
      "keywords": "data manifold;manifold learning;generalization bounds  controlled datasets  deep learning theory",
      "tldr": "",
      "primary_area": "learning theory",
      "similarity_score": 10.449340539110574,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.449340539110574,
      "combined_score": 0.0,
      "rank": 90
    },
    {
      "title": "Geo-NN: An End-to-End Framework for Geodesic Mean Estimation on the Manifold of Symmetric Positive Definite Matrices",
      "authors": [
        "Niharika Shimona D'Souza",
        "Archana Venkataraman"
      ],
      "abstract": "The manifold of symmetric positive definite (SPD) matrices plays a key role in many domains, from network science to differential geometry to signal and image processing. However, leveraging the SPD manifold geometry during inference is challenging, as simple operations, such as mean estimation, do not have a closed-form or easily computable solution. In this paper, we propose an end-to-end deep learning framework, which we call a Geometric Neural Network (Geo-NN), to efficiently compute the geodesic mean of a collection of matrices lying on the SPD manifold. Geo-NN utilizes a Matrix-Autoencoder (MAE) architecture with intersecting fully connected layers as its backbone. We illustrate that the matrix-normal equation arising from Fr\\'echet mean estimation can be converted into a loss function for optimizing the Geo-NN, which in turn approximates the geodesic mean of a collection of SPD matrices. We demonstrate the efficacy of our framework in both synthetic and real-world scenarios, as compared to commonly used alternative methods. Our simulated experiments demonstrate that Geo-NN is robust to various noise conditions and is scalable to increasing dataset size and dimensionality. Our real-world application of Geo-NN to functional connectomics data allows us to extract network patterns associated with patient/control differences.",
      "url": "https://openreview.net/forum?id=h-UkhDzFFj",
      "year": 2023,
      "venue": "ICLR 2023",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "h-UkhDzFFj",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Symmetric Postive Definite Manifolds;Geodesic Mean;Matrix Autoencoder",
      "tldr": "We propose an end-to-end deep learning framework, the Geo-NN, to efficiently compute the geodesic mean of a collection of matrices lying on the SPD manifold",
      "primary_area": "",
      "similarity_score": 10.402290379358275,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.402290379358275,
      "combined_score": 0.0,
      "rank": 91
    },
    {
      "title": "Understanding Scaling Laws with Statistical and Approximation Theory for Transformer Neural Networks on Intrinsically Low-dimensional Data",
      "authors": [
        "Alexander Havrilla",
        "Wenjing Liao"
      ],
      "abstract": "When training deep neural networks, a model's generalization error is often observed to follow a power scaling law dependent both on the model size and the data size. Perhaps the best known example of such scaling laws are for transformer-based large language models (**LLMs**), where networks with billions of parameters are trained on trillions of tokens of text. Yet, despite sustained widespread interest, a rigorous understanding of why transformer scaling laws exist is still missing. To answer this question, we establish novel statistical estimation and mathematical approximation theories for transformers when the input data are concentrated on a low-dimensional manifold. Our theory predicts a power law between the generalization error and both the training data size and the network size for transformers, where the power depends on the intrinsic dimension $d$ of the training data. Notably, the constructed model architecture is shallow, requiring only logarithmic depth in $d$. By leveraging low-dimensional data structures under a manifold hypothesis, we are able to explain transformer scaling laws in a way which respects the data geometry.  Moreover, we test our theory with empirical observation by training LLMs on natural language datasets. We find the observed empirical scaling laws closely agree with our theoretical predictions. Taken together, these results rigorously show the intrinsic dimension of data to be a crucial quantity affecting transformer scaling laws in both theory and practice.",
      "url": "https://neurips.cc/virtual/2024/poster/95466",
      "year": 2024,
      "venue": "NIPS 2024",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=N2wYPMpifA",
      "citations": null,
      "categories": [],
      "id": "N2wYPMpifA",
      "track": "main",
      "status": "Poster",
      "keywords": "scaling laws;LLMs;approximation theory;statistical theory",
      "tldr": "",
      "primary_area": "learning_theory",
      "similarity_score": 10.38319654715644,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.38319654715644,
      "combined_score": 0.0,
      "rank": 92
    },
    {
      "title": "LoRA-S: An Efficient Low Rank Adaptation scheme via Sylvester equation",
      "authors": [],
      "abstract": "Numerous studies on low-rank adaptation (LoRA) emerged in recent years, with the aim of accelerating the convergence of the LoRA framework. In this paper, we leverage the horizontal lift theory from differential geometry to establish the general iteration scheme on the quotient manifold  \\mathbb{R}\\_\\*^{m \\times r} \\times \\mathbb{R}\\_\\*^{n \\times r}/\\sim. \nBy endowing the LoRA framework with Riemannian quotient geometries, our theory not only guarantees efficient feature learning but also bridges the LoRA algorithms and the pre-training algorithms for large models. \nFurthermore, we theoretically analyze the role of the weight decay matrix $\\epsilon_{decay}I$ in efficient feature learning and then replace it with the Sylvester matrix $K$, indicating that the theory helps remove an important hyperparameter while generating accurate and computationally efficient optimizers. \nBased on the general scheme, we propose two efficient LoRA optimizers with runtime analysis, Adam-Sylvester (AdamS) and LRACS, then conduct experiments on the transformer-based networks. The results demonstrate evident improvements over existing optimizers.",
      "url": "https://openreview.net/forum?id=Guo2XGgxZA",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "Guo2XGgxZA",
      "track": "main",
      "status": "Active",
      "keywords": "optimization;LoRA",
      "tldr": "",
      "primary_area": "optimization",
      "similarity_score": 10.373421137070343,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.373421137070343,
      "combined_score": 0.0,
      "rank": 93
    },
    {
      "title": "Manifold K-means with $\\ell_{2,p}$-Norm Maximization",
      "authors": [
        "Fangfang Li",
        "Quanxue Gao",
        "Qianqian Wang",
        "Cheng Deng",
        "Xiaoke Ma",
        "Jing Li"
      ],
      "abstract": "Although a variety of different methods have emerged in the field of clustering, K-means still occupies an important position, and many advanced clustering methods even rely on the K-means  to achieve effective cluster detection. However, the sensitivity of K-means to the selection of the initial cluster center and its limited ability to handle nonlinear separable data somewhat restrict its clustering performance. In order to overcome the limitations of K-means, we draw inspiration from manifold learning and redefine K-means as a manifold K-means clustering framework. This framework supports various types of distance matrices, thus facilitating the efficient processing of nonlinear separable data. A unique advantage of this approach is that it does not require the calculation of the cluster center, while it maintains the consistency between manifold structure and cluster labels. Additionally, we highlight the significant role of the $\\ell_{2,p}$-norm; by maximizing the $\\ell_{2,p}$-norm, we can ensure the balance of classes in the clustering process, which is also supported by theoretical analysis. The results from extensive experiments across multiple databases substantiate the superiority of our proposed model.",
      "url": "https://openreview.net/forum?id=E5DYpUWsES",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "E5DYpUWsES",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Clustering;Manifold Learning;K-means;$\\ell_{2;p}$-Norm",
      "tldr": "",
      "primary_area": "unsupervised, self-supervised, semi-supervised, and supervised representation learning",
      "similarity_score": 10.347451350469136,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.347451350469136,
      "combined_score": 0.0,
      "rank": 94
    },
    {
      "title": "Discrete Probabilistic Inverse Optimal Transport",
      "authors": [
        "Wei-Ting Chiu",
        "Pei Wang",
        "Patrick Shafto"
      ],
      "abstract": "Inverse Optimal Transport (IOT) studies the problem of inferring the underlying cost that gives rise to an observation on coupling two probability measures. Couplings appear as the outcome of matching sets (e.g. dating) and moving distributions (e.g. transportation). Compared to Optimal transport (OT), the mathematical theory of IOT is undeveloped. We formalize and systematically analyze the properties of IOT using tools from the study of entropy-regularized OT. Theoretical contributions include characterization of the manifold of cross-ratio equivalent costs, the implications of model priors, and derivation of an MCMC sampler. Empirical contributions include visualizations of cross-ratio equivalent effect on basic examples, simulations validating theoretical results and experiments on real world data.",
      "url": "https://icml.cc/virtual/2022/poster/16637",
      "year": 2022,
      "venue": "ICML 2022",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://proceedings.mlr.press/v162/chiu22b/chiu22b.pdf",
      "citations": null,
      "categories": [],
      "id": "16637",
      "track": "main",
      "status": "Spotlight",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.309493553943334,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.309493553943334,
      "combined_score": 0.0,
      "rank": 95
    },
    {
      "title": "Local Identifiability of Deep ReLU Neural Networks: the Theory",
      "authors": [
        "Joachim Bona-Pellissier",
        "Francois Malgouyres",
        "Francois Bachoc"
      ],
      "abstract": "Is a sample rich enough to determine, at least locally, the parameters of a neural network? To answer this question, we introduce a new local parameterization of a given deep ReLU neural network by fixing the values of some of its weights. This allows us to define local lifting operators whose inverses are charts of a smooth manifold of a high dimensional space. The function implemented by the deep ReLU neural network composes the local lifting with a linear operator which depends on the sample. We derive from this convenient representation a geometrical necessary and sufficient condition of local identifiability. Looking at tangent spaces, the geometrical condition provides: 1/ a sharp and testable necessary condition of identifiability and 2/ a sharp and testable sufficient condition of local identifiability. The validity of the conditions can be tested numerically using backpropagation and matrix rank computations.",
      "url": "https://nips.cc/virtual/2022/poster/53394",
      "year": 2022,
      "venue": "NIPS 2022",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=-3cHWtrbLYq",
      "citations": null,
      "categories": [],
      "id": "-3cHWtrbLYq",
      "track": "main",
      "status": "Accept",
      "keywords": "Deep Learning;ReLU networks;Conditions of identifiability;Lifting operator",
      "tldr": "We characterize theoretically the question of local identifiability for deep ReLU neural networks and we provide numerically testable conditions.",
      "primary_area": "",
      "similarity_score": 10.296487815192744,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.296487815192744,
      "combined_score": 0.0,
      "rank": 96
    },
    {
      "title": "Low-rank tensor completion: a Riemannian manifold preconditioning approach",
      "authors": [
        "Hiroyuki Kasai",
        "Bamdev Mishra"
      ],
      "abstract": "We propose a novel Riemannian manifold preconditioning approach for the tensor completion problem with rank constraint. A novel Riemannian metric or inner product is proposed that exploits the least-squares structure of the cost function and takes into account the structured symmetry that exists in Tucker decomposition. The specific metric allows to use the versatile framework of Riemannian optimization on quotient manifolds to develop preconditioned nonlinear conjugate gradient and stochastic gradient descent algorithms in batch and online setups, respectively. Concrete matrix representations of various optimization-related ingredients are listed. Numerical comparisons suggest that our proposed algorithms robustly outperform state-of-the-art algorithms across different synthetic and real-world datasets.",
      "url": "https://proceedings.mlr.press/v48/kasai16.html",
      "year": 2016,
      "venue": "ICML 2016",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "http://proceedings.mlr.press/v48/kasai16.pdf",
      "citations": null,
      "categories": [],
      "id": "38fdde6038",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.252272825760722,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.252272825760722,
      "combined_score": 0.0,
      "rank": 97
    },
    {
      "title": "Anomaly Detection Based on Unsupervised Disentangled Representation Learning in Combination with Manifold Learning",
      "authors": [
        "Xiaoyan Li",
        "Iluju Kiringa",
        "Tet Yeap",
        "Xiaodan Zhu",
        "Yifeng Li"
      ],
      "abstract": "Identifying anomalous samples from highly complex and unstructured data is a crucial but challenging task in a variety of intelligent systems. In this paper, we present a novel deep anomaly detection framework named AnoDM (standing for Anomaly detection based on unsupervised Disentangled representation learning and Manifold learning). The disentanglement learning is currently implemented by beta-VAE for automatically discovering interpretable factorized latent representations in a completely unsupervised manner. The manifold learning is realized by t-SNE for projecting the latent representations to a 2D map.  We define a new anomaly score function by combining beta-VAE's reconstruction error in the raw feature space and local density estimation in the t-SNE space. AnoDM was evaluated on both image and time-series data and achieved better results than models that use just one of the two measures and other deep learning methods.",
      "url": "https://openreview.net/forum?id=r1xHxgrKwr",
      "year": 2020,
      "venue": "ICLR 2020",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "r1xHxgrKwr",
      "track": "main",
      "status": "Reject",
      "keywords": "anomaly detection;disentangled representation learning;manifold learning",
      "tldr": "We developed anomaly detection framework based on beta-VAE and t-SNE",
      "primary_area": "",
      "similarity_score": 10.233999176865328,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.233999176865328,
      "combined_score": 0.0,
      "rank": 98
    },
    {
      "title": "SMaRt: Improving GANs with Score Matching Regularity",
      "authors": [
        "Mengfei Xia",
        "Yujun Shen",
        "Ceyuan Yang",
        "Ran Yi",
        "Wenping Wang",
        "Yong-jin Liu"
      ],
      "abstract": "Generative adversarial networks (GANs) usually struggle in learning from highly diverse data, whose underlying manifold is complex. In this work, we revisit the mathematical foundations of GANs, and theoretically reveal that the native adversarial loss for GAN training is insufficient to fix the problem of $\\textit{subsets with positive Lebesgue measure of the generated data manifold lying out of the real data manifold}$. Instead, we find that score matching serves as a promising solution to this issue thanks to its capability of persistently pushing the generated data points towards the real data manifold. We thereby propose to improve the optimization of GANs with score matching regularity (SMaRt). Regarding the empirical evidences, we first design a toy example to show that training GANs by the aid of a ground-truth score function can help reproduce the real data distribution more accurately, and then confirm that our approach can consistently boost the synthesis performance of various state-of-the-art GANs on real-world datasets with pre-trained diffusion models acting as the approximate score function. For instance, when training Aurora on the ImageNet $64\\times64$ dataset, we manage to improve FID from 8.87 to 7.11, on par with the performance of one-step consistency model. Code is available at https://github.com/thuxmf/SMaRt.",
      "url": "https://icml.cc/virtual/2024/poster/33200",
      "year": 2024,
      "venue": "ICML 2024",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=lqeVCc9zYq",
      "citations": null,
      "categories": [],
      "id": "lqeVCc9zYq",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.230722939914116,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.230722939914116,
      "combined_score": 0.0,
      "rank": 99
    },
    {
      "title": "Manifold Preserving Guided Diffusion",
      "authors": [
        "Yutong He",
        "Naoki Murata",
        "Chieh-Hsin Lai",
        "Yuhta Takida",
        "Toshimitsu Uesaka",
        "Dongjun Kim",
        "Wei-Hsiang Liao",
        "Yuki Mitsufuji",
        "J Zico Kolter",
        "Ruslan Salakhutdinov"
      ],
      "abstract": "Despite the recent advancements, conditional image generation still faces challenges of cost, generalizability, and the need for task-specific training. In this paper, we propose Manifold Preserving Guided Diffusion (MPGD), a training-free conditional generation framework that leverages pretrained diffusion models and off-the-shelf neural networks with minimal additional inference cost for a broad range of tasks. Specifically, we leverage the manifold hypothesis to refine the guided diffusion steps and introduce a shortcut algorithm in the process. We then propose two methods for on-manifold training-free guidance using pre-trained autoencoders and demonstrate that our shortcut inherently preserves the manifolds when applied to latent diffusion models. Our experiments show that MPGD is efficient and effective for solving a variety of conditional generation applications in low-compute settings, and can consistently offer up to 3.8× speed-ups with the same number of diffusion steps while maintaining high sample quality compared to the baselines.",
      "url": "https://iclr.cc/virtual/2024/poster/17837",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=o3BxOLoxm1",
      "citations": null,
      "categories": [],
      "id": "o3BxOLoxm1",
      "track": "main",
      "status": "Poster",
      "keywords": "generative model;diffusion model;controllable generation",
      "tldr": "",
      "primary_area": "generative models",
      "similarity_score": 10.220907561140773,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.220907561140773,
      "combined_score": 0.0,
      "rank": 100
    },
    {
      "title": "Manifold Topology Divergence: a Framework for Comparing Data Manifolds.",
      "authors": [
        "Serguei Barannikov",
        "Ilya Trofimov",
        "Grigorii Sotnikov",
        "Ekaterina Trimbach",
        "Alexander Korotin",
        "Alexander Filippov",
        "Evgeny Burnaev"
      ],
      "abstract": "We propose a framework for comparing data manifolds, aimed, in particular, towards the evaluation of deep generative models. We describe a novel tool, Cross-Barcode(P,Q), that, given a pair of distributions in a high-dimensional space, tracks multiscale topology spacial discrepancies between manifolds on which the distributions are concentrated. Based on the Cross-Barcode, we introduce the Manifold Topology Divergence score (MTop-Divergence) and apply it to assess the performance of deep generative models in various domains: images, 3D-shapes, time-series, and on different datasets: MNIST, Fashion MNIST, SVHN, CIFAR10, FFHQ, market stock data, ShapeNet. We demonstrate that the MTop-Divergence accurately detects various degrees of mode-dropping, intra-mode collapse, mode invention, and image disturbance. Our algorithm scales well (essentially linearly) with the increase of the dimension of the ambient high-dimensional space. It is one of the first TDA-based methodologies that can be applied universally to datasets of different sizes and dimensions, including the ones on which the most recent GANs in the visual domain are trained. The proposed method is domain agnostic and does not rely on pre-trained networks.",
      "url": "https://nips.cc/virtual/2021/poster/27062",
      "year": 2021,
      "venue": "NIPS 2021",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=Fj6kQJbHwM9",
      "citations": null,
      "categories": [],
      "id": "Fj6kQJbHwM9",
      "track": "main",
      "status": "Poster",
      "keywords": "data manifolds;point clouds;persistent homology;topology;generative models;generative adversarial networks;mode-dropping;3D-shapes;time-series",
      "tldr": "We introduce a topology-based domain agnostic methodology for comparing data manifolds.",
      "primary_area": "",
      "similarity_score": 10.199562118370032,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.199562118370032,
      "combined_score": 0.0,
      "rank": 101
    },
    {
      "title": "Language Guided Interpretable Image Recognition via Manifold Alignment",
      "authors": [
        "Jiaqi Wang",
        "Pichao WANG",
        "Fan Wang",
        "Liping Jing"
      ],
      "abstract": "Most works of interpretable neural networks strive for learning the semantics concepts merely from single modal information such as images. However, humans usually learn semantic concepts from multiple modalities and the semantics is encoded by the brain from fused multi-modal information. Inspired by cognitive science and vision-language learning, we propose a two-stream model for learning visual semantic concepts under the guidance of natural language, where a CNN-based vision stream encodes the input image and a Bert-based language stream encodes corresponding text description. Therefore, visual and natural language features reside on different but semantically highly correlated manifolds, \\ie follow a multi-manifold distribution.  We transform the multi-manifold distribution alignment problem into updating the projection matrices by Cayley transform on the Stiefel manifold and  better joint representations are obtained by fusing the semantically similar features from the aligned manifold. In addition, we propose a Manifold Alignment based Prototypical Part Network (MA-ProtoPNet) to learn the semantics concepts from the joint representations, and these concepts can capture more semantic information from multi-modality. We verified the effectiveness of the manifold alignment method through experiments and the proposed framework can provide better interpretability and classification results.",
      "url": "https://openreview.net/forum?id=Cn9Cl08zSS",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "Cn9Cl08zSS",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Explainable AI;Prototypes;Manifold Alignment",
      "tldr": "",
      "primary_area": "representation learning for computer vision, audio, language, and other modalities",
      "similarity_score": 10.167218999705234,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.167218999705234,
      "combined_score": 0.0,
      "rank": 102
    },
    {
      "title": "Canonical normalizing flows for manifold learning",
      "authors": [
        "Kyriakos Flouris",
        "Ender Konukoglu"
      ],
      "abstract": "Manifold learning flows are a class of generative modelling techniques that assume a low-dimensional manifold description of the data. The embedding of such a manifold into the high-dimensional space of the data is achieved via learnable invertible transformations. Therefore, once the manifold is properly aligned via a reconstruction loss, the probability density is tractable on the manifold and maximum likelihood can be used to optimize the network parameters. Naturally, the lower-dimensional representation of the data requires an injective-mapping. Recent approaches were able to enforce that the density aligns with the modelled manifold, while efficiently calculating the density volume-change term when embedding to the higher-dimensional space. However, unless the injective-mapping is analytically predefined, the learned manifold is not necessarily an \\emph{efficient representation} of the data. Namely, the latent dimensions of such models frequently learn an entangled intrinsic basis, with degenerate information being stored in each dimension. Alternatively, if a locally orthogonal and/or sparse basis is to be learned, here coined canonical intrinsic basis, it can serve in learning a more compact latent space representation. Toward this end, we propose a canonical manifold learning flow method, where a novel optimization objective enforces the transformation matrix to have few prominent and non-degenerate basis functions. We demonstrate that by minimizing the off-diagonal manifold metric elements $\\ell_1$-norm, we can achieve such a basis, which is simultaneously sparse and/or orthogonal. Canonical manifold flow yields a more efficient use of the latent space, automatically generating fewer prominent and distinct dimensions to represent data, and consequently a better approximation of target distributions than other manifold flow methods in most experiments we conducted, resulting in lower FID scores.",
      "url": "https://nips.cc/virtual/2023/poster/69924",
      "year": 2023,
      "venue": "NIPS 2023",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=yubwSWol6K",
      "citations": null,
      "categories": [],
      "id": "yubwSWol6K",
      "track": "main",
      "status": "Poster",
      "keywords": "manifold learning flows;normalizing flows;optimization;orthogonalization;sparsity;sparse learning;generative modeling;Riemannian manifold;geometry;metric tensor;orthogonal basis",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.079718747414415,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.079718747414415,
      "combined_score": 0.0,
      "rank": 103
    },
    {
      "title": "Manifold Learning Benefits GANs",
      "authors": [
        "Yao Ni",
        "Piotr Koniusz",
        "Richard Hartley",
        "Richard Nock"
      ],
      "abstract": "In this paper, we improve Generative Adversarial Networks by incorporating a manifold learning step into the discriminator. We consider locality-constrained linear and subspace-based manifolds, and locality-constrained non-linear manifolds. In our design, the manifold learning and coding steps are intertwined with layers of the discriminator, with the goal of attracting intermediate feature representations onto manifolds. We adaptively balance the discrepancy between feature representations and their manifold view, which is a trade-off between denoising on the manifold and refining the manifold. We find that locality-constrained non-linear manifolds outperform linear manifolds due to their non-uniform density and smoothness. We also substantially outperform state-of-the-art baselines.",
      "url": "",
      "year": 2022,
      "venue": "CVPR 2022",
      "source": "offline_cvpr",
      "doi": null,
      "pdf_url": "https://openaccess.thecvf.com/content/CVPR2022/papers/Ni_Manifold_Learning_Benefits_GANs_CVPR_2022_paper.pdf",
      "citations": null,
      "categories": [],
      "id": "",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 10.040466916840574,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 10.040466916840574,
      "combined_score": 0.0,
      "rank": 104
    },
    {
      "title": "Information Diffusion Kernels",
      "authors": [
        "Guy Lebanon",
        "John D. Lafferty"
      ],
      "abstract": "A new family of kernels for statistical learning is introduced that ex- ploits the geometric structure of statistical models. Based on the heat equation on the Riemannian manifold defined by the Fisher informa- tion metric, information diffusion kernels generalize the Gaussian kernel of Euclidean space, and provide a natural way of combining generative statistical modeling with non-parametric discriminative learning. As a special case, the kernels give a new approach to applying kernel-based learning algorithms to discrete data. Bounds on covering numbers for the new kernels are proved using spectral theory in differential geometry, and experimental results are presented for text classification.",
      "url": "https://papers.nips.cc/paper_files/paper/2002/hash/5938b4d054136e5d59ada6ec9c295d7a-Abstract.html",
      "year": 2002,
      "venue": "NIPS 2002",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2002/file/5938b4d054136e5d59ada6ec9c295d7a-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "11597ce1e1",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 9.945882968341586,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 9.945882968341586,
      "combined_score": 0.0,
      "rank": 105
    },
    {
      "title": "Information-Ordered Bottlenecks for Adaptive Dimensionality Reduction",
      "authors": [
        "Matthew Ho",
        "Xiaosheng Zhao",
        "Benjamin Dan Wandelt"
      ],
      "abstract": "We present the information-ordered bottleneck (IOB), a neural layer designed to adaptively compress data into latent variables ordered by likelihood maximization. Without retraining, IOB nodes can be truncated at any bottleneck width, capturing the most crucial information in the first latent variables. Unifying several previous approaches, we show that IOBs achieve near-optimal compression for a given encoding architecture and can assign ordering to latent signals in a manner that is semantically meaningful. IOBs demonstrate a remarkable ability to compress embeddings of image and text data, leveraging the performance of SOTA architectures such as CNNs, transformers, and diffusion models. Moreover, we introduce a novel theory for estimating global intrinsic dimensionality with IOBs and show that they recover SOTA dimensionality estimates for complex synthetic data. Furthermore, we showcase the utility of these models for exploratory analysis through applications on heterogeneous datasets, enabling computer-aided discovery of dataset complexity.",
      "url": "https://openreview.net/forum?id=XbydvPq92M",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "XbydvPq92M",
      "track": "main",
      "status": "Reject",
      "keywords": "Deep Learning;Nonlinear Dimension Reduction and Manifold Learning;Neural Networks;Component Analysis (ICA;PCA;CCA;FLDA);Compressed Sensing and Sparse Reconstruction",
      "tldr": "",
      "primary_area": "unsupervised, self-supervised, semi-supervised, and supervised representation learning",
      "similarity_score": 9.91238992348412,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 9.91238992348412,
      "combined_score": 0.0,
      "rank": 106
    },
    {
      "title": "Adaptive Manifold Learning",
      "authors": [
        "Jing Wang",
        "Zhenyue Zhang",
        "Hongyuan Zha"
      ],
      "abstract": "Recently, there have been several advances in the machine learning and           pattern recognition communities for developing manifold learning algo-           rithms to construct nonlinear low-dimensional manifolds from sample           data points embedded in high-dimensional spaces. In this paper, we de-           velop algorithms that address two key issues in manifold learning: 1)           the adaptive selection of the neighborhood sizes; and 2) better fitting the           local geometric structure to account for the variations in the curvature           of the manifold and its interplay with the sampling density of the data           set. We also illustrate the effectiveness of our methods on some synthetic           data sets.",
      "url": "https://papers.nips.cc/paper_files/paper/2004/hash/eb0ecdb070a1a0ac46de0cd733d39cf3-Abstract.html",
      "year": 2004,
      "venue": "NIPS 2004",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2004/file/eb0ecdb070a1a0ac46de0cd733d39cf3-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "2d2f902017",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 9.879745255380424,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 9.879745255380424,
      "combined_score": 0.0,
      "rank": 107
    },
    {
      "title": "The Gyro-Structure of Some Matrix Manifolds",
      "authors": [
        "Xuan Son Nguyen"
      ],
      "abstract": "In this paper, we study the gyrovector space structure (gyro-structure) of matrix manifolds. Our work is motivated by the success of hyperbolic neural networks (HNNs) that have demonstrated impressive performance in a variety of applications. At the heart of HNNs is the theory of gyrovector spaces that provides a powerful tool for studying hyperbolic geometry. Here we focus on two matrix manifolds, i.e., Symmetric Positive Definite (SPD) and Grassmann manifolds, and consider connecting the Riemannian geometry of these manifolds with the basic operations, i.e., the binary operation and scalar multiplication on gyrovector spaces. Our work reveals some interesting facts about SPD and Grassmann manifolds. First, SPD matrices with an Affine-Invariant (AI) or a Log-Euclidean (LE) geometry have rich structure with strong connection to hyperbolic geometry. Second, linear subspaces, when equipped with our proposed basic operations, form what we call gyrocommutative and gyrononreductive gyrogroups. Furthermore, they share remarkable analogies with gyrovector spaces. We demonstrate the applicability of our approach for human activity understanding and question answering.",
      "url": "https://nips.cc/virtual/2022/poster/53244",
      "year": 2022,
      "venue": "NIPS 2022",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=eyE9Fb2AvOT",
      "citations": null,
      "categories": [],
      "id": "eyE9Fb2AvOT",
      "track": "main",
      "status": "Accept",
      "keywords": "manifold learning;representation learning;deep learning;gyrovector spaces",
      "tldr": "This paper studies the gyrovector space structure (gyro-structure) of some matrix manifolds",
      "primary_area": "",
      "similarity_score": 9.876371761792559,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 9.876371761792559,
      "combined_score": 0.0,
      "rank": 108
    },
    {
      "title": "Flows for simultaneous manifold learning and density estimation",
      "authors": [
        "Johann Brehmer",
        "Kyle Cranmer"
      ],
      "abstract": "We introduce manifold-learning flows (ℳ-flows), a new class of generative models that simultaneously learn the data manifold as well as a tractable probability density on that manifold. Combining aspects of normalizing flows, GANs, autoencoders, and energy-based models, they have the potential to represent data sets with a manifold structure more faithfully and provide handles on dimensionality reduction, denoising, and out-of-distribution detection. We argue why such models should not be trained by maximum likelihood alone and present a new training algorithm that separates manifold and density updates. In a range of experiments we demonstrate how ℳ-flows learn the data manifold and allow for better inference than standard flows in the ambient data space.",
      "url": "https://nips.cc/virtual/2020/poster/17120",
      "year": 2020,
      "venue": "NIPS 2020",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2020/file/051928341be67dcba03f0e04104d9047-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "17120",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 9.861421352659839,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 9.861421352659839,
      "combined_score": 0.0,
      "rank": 109
    },
    {
      "title": "Riemannian Low-Rank Adaptation for Federated Fine-Tuning of Foundation Models",
      "authors": [
        "Zihan Zhou",
        "Yang Zhou",
        "Tianshi Che",
        "Zeru Zhang",
        "Jiaxiang Ren",
        "Da Yan",
        "Zhe Jiang",
        "yelong shen",
        "Ruoming Jin",
        "Jianfeng Gao"
      ],
      "abstract": "Rank-adaptive low-rank adaptation (LoRA), a parameter-efficient fine-tuning (PEFT) technology, has achieved state-of-the-art performance in fine-tuning foundation models (FM). Directly transplanting the rank-adaptive LoRA methods from centralized learning to federated learning raises two critical issues: client drift and rank drift. This paper presents a Riemannian LoRA algorithm with adaptive rank for federated fine-tuning of foundation models (FFT-FM), RAFFT, which solves the client-drift and rank-drift issues, and significantly improves the computational cost. First, by utilizing Riemannian Procrustes analysis, we propose a Riemannian parameter matching method to avoid the client-drift issue for ensuring the effectiveness of FFT-FM with rank-adaptive LoRA, and to reduce the cost of matrix decomposition by transforming the singular value decomposition (SVD) of high-dimensional full parameter matrices into the SVD of low-dimensional $r \\times r$ matrices, where $r$ is the rank parameter in the LoRA. We theoretically derive the equivalence between our RAFFT algorithm with rank-adaptive LoRA for the FFT-FM and the standard FFT-FM on the full parameter matrices based on FedAvg and verify the bounded error introduced by approximation and numerical errors. Second, by leveraging Riemannian manifold theory, we develop a Riemannian gradient descent (RGD) method to guarantee the local full parameter matrices on clients in the form of low-rank ones with fixed rank optimized by the server in each FFT-FM round, for alleviating the rank-drift issue to speed up the convergence of RAFFT. We theoretically demonstrate that the RGD optimization on the Riemannian manifold ensures the rank invariance during the local update process and the RGD optimization can converge in the FFT-FM context.",
      "url": "https://openreview.net/forum?id=lbasmwFWzH",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "lbasmwFWzH",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Rank-adaptive LoRA;Federated Learning;Fine-Tuning;Foundation Models;Riemannian Theory",
      "tldr": "",
      "primary_area": "other topics in machine learning (i.e., none of the above)",
      "similarity_score": 9.856540947788265,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 9.856540947788265,
      "combined_score": 0.0,
      "rank": 110
    },
    {
      "title": "Content Moderation and the Formation of Online Communities: A Theoretical Framework",
      "authors": [
        "Cynthia Dwork",
        "Chris Hays",
        "Jon Kleinberg",
        "Manish Raghavan"
      ],
      "abstract": "",
      "url": "",
      "year": 2024,
      "venue": "WWW 2024",
      "source": "offline_www",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "aDjVZYOZzR",
      "track": "main",
      "status": "Oral",
      "keywords": "content moderation;online platforms;online communities;social media",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 7.647435789496325,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.647435789496325,
      "combined_score": 0.0,
      "rank": 111
    },
    {
      "title": "A General Theoretical Framework for Learning Smallest Interpretable Models",
      "authors": [
        "Sebastian Ordyniak",
        "Giacomo Paesani",
        "Mateusz Rychlicki",
        "Stefan Szeider"
      ],
      "abstract": "We develop a general algorithmic framework that allows us to obtain fixed-parameter tractability for computing smallest symbolic models that represent given data. Our framework applies to all ML model types that admit a certain extension property. By showing this extension property for decision trees, decision sets, decision lists, and binary decision diagrams, we obtain that minimizing these fundamental model types is fixed-parameter tractable. Our framework even applies to ensembles, which combine individual models by majority decision.",
      "url": "https://ojs.aaai.org/index.php/AAAI/article/view/28937",
      "year": 2024,
      "venue": "AAAI 2024",
      "source": "offline_aaai",
      "doi": null,
      "pdf_url": "https://ojs.aaai.org/index.php/AAAI/article/view/28937/29781",
      "citations": null,
      "categories": [],
      "id": "article-28937",
      "track": "main",
      "status": "Technical",
      "keywords": "",
      "tldr": "",
      "primary_area": "knowledge representation and reasoning",
      "similarity_score": 7.31331635834114,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.31331635834114,
      "combined_score": 0.0,
      "rank": 112
    },
    {
      "title": "CoT-Space: A Theoretical Framework for Internal Slow-Thinking via Reinforcement Learning",
      "authors": [],
      "abstract": "Reinforcement Learning (RL) has become a pivotal approach for enhancing the reasoning capabilities of Large Language Models (LLMs). However, a significant theoretical gap persists, as traditional token-level RL frameworks fail to align with the reasoning-level nature of complex, multi-step thought processes like Chain-of-Thought (CoT). To address this challenge, we introduce CoT-Space, a novel theoretical framework that recasts LLM reasoning from a discrete token-prediction task to an optimization process within a continuous, reasoning-level semantic space. This shift in perspective serves as a conceptual bridge, revitalizing foundational principles from classical learning theory to analyze the unique dynamics of LLMs. By analyzing this process from both a noise perspective and a risk perspective, we demonstrate that the convergence to an optimal CoT length is a natural consequence of the fundamental trade-off between underfitting and overfitting. Furthermore, extensive experiments provide strong empirical validation for our theoretical findings. Our framework not only provides a coherent explanation for empirical phenomena such as overthinking but also offers a solid theoretical foundation to guide the future development of more effective and generalizable reasoning agents. We open-source our code through an anonymous GitHub repository at https://anonymous.4open.science/r/CoT-Space-Reasoning-via-RL.",
      "url": "https://openreview.net/forum?id=4DJ62eO2T2",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "4DJ62eO2T2",
      "track": "main",
      "status": "Active",
      "keywords": "large language model;reasoning;test-time scaling",
      "tldr": "",
      "primary_area": "foundation or frontier models, including LLMs",
      "similarity_score": 7.303893318732209,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.303893318732209,
      "combined_score": 0.0,
      "rank": 113
    },
    {
      "title": "A Theoretical Framework for Preventing Class Collapse in Supervised Contrastive Learning",
      "authors": [
        "Chungpa Lee",
        "Jeongheon Oh",
        "Kibok Lee",
        "Jy-yong Sohn"
      ],
      "abstract": "Supervised contrastive learning (SupCL) has emerged as a prominent approach in representation learning, leveraging both supervised and self-supervised losses. However, achieving an optimal balance between these losses is challenging; failing to do so can lead to class collapse, reducing discrimination among individual embeddings in the same class. In this paper, we present theoretically grounded guidelines for SupCL to prevent class collapse in learned representations. Specifically, we introduce the Simplex-to-Simplex Embedding Model (SSEM), a theoretical framework that models various embedding structures, including all embeddings that minimize the supervised contrastive loss. Through SSEM, we analyze how hyperparameters affect learned representations, offering practical guidelines for hyperparameter selection to mitigate the risk of class collapse. Our theoretical findings are supported by empirical results across synthetic and real-world datasets.",
      "url": "https://openreview.net/forum?id=ElvhiUFA02",
      "year": 2025,
      "venue": "AISTATS 2025",
      "source": "offline_aistats",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "ElvhiUFA02",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 7.257290443941222,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.257290443941222,
      "combined_score": 0.0,
      "rank": 114
    },
    {
      "title": "A Causal Theoretical Framework for Open Set Domain Adaptation",
      "authors": [
        "Huaming Du",
        "Lei Yuan",
        "Gang Kou",
        "Carl Yang"
      ],
      "abstract": "Open Set Domain Adaptation (OSDA) faces two critical challenges: the emergence\nof unknown classes in the target domain and changes in observed distributions\nacross domains. Although numerous studies have proposed advanced algorithms,\nrecent experimental results demonstrate that the classical Empirical Risk Mini\u0002mization (ERM) approach still delivers state-of-the-art performance. However,\nfew theories can effectively explain this disputed phenomenon. To address the\ntheoretical gap, we focus on constructing a causal theoretical framework for OSDA.\nWe formulate the novel concepts of the Fully Informative Causal Invariance Model\n(FICIM) and the Partially Informative Causal Invariance Model (PICIM). Subse\u0002quently, We derive an OSDA theoretical bound to prove that the ERM performs\nwell when the source domain follows FICIM, while it performs poorly when the\nsource domain follows PICIM. The different results may be attributed to the vary\u0002ing amounts of available information when bounding the target domain’s stable\nexpected risk. Finally, across different datasets, we conduct extensive experiments\non the FICIM and PICIM source domains to validate the effectiveness of our\ntheoretical results.",
      "url": "https://openreview.net/forum?id=Gp6VU0oJX3",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "Gp6VU0oJX3",
      "track": "main",
      "status": "Reject",
      "keywords": "Causal theory;open set domain adaptation;domain adaptation;empirical risk minimization",
      "tldr": "",
      "primary_area": "causal reasoning",
      "similarity_score": 7.253724642694396,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.253724642694396,
      "combined_score": 0.0,
      "rank": 115
    },
    {
      "title": "Balancing the Scales: A Theoretical and Algorithmic Framework for Learning from Imbalanced Data",
      "authors": [
        "Corinna Cortes",
        "Anqi Mao",
        "Mehryar Mohri",
        "Yutao Zhong"
      ],
      "abstract": "Class imbalance remains a major challenge in machine learning, especially in multi-class problems with long-tailed distributions. Existing methods, such as data resampling, cost-sensitive techniques, and logistic loss modifications, though popular and often effective, lack solid theoretical foundations. As an example, we demonstrate that cost-sensitive methods  are not Bayes-consistent. This paper introduces a novel theoretical framework for analyzing generalization in imbalanced classification. We propose a new class-imbalanced margin loss function for both binary and multi-class settings, prove its strong $H$-consistency, and derive corresponding learning guarantees based on empirical loss and a new notion of class-sensitive Rademacher complexity. Leveraging these theoretical results, we devise novel and general learning algorithms, IMMAX (*Imbalanced Margin Maximization*), which incorporate confidence margins and are applicable to various hypothesis sets. While our focus is theoretical, we also present extensive empirical results demonstrating the effectiveness of our algorithms compared to existing baselines.",
      "url": "https://icml.cc/virtual/2025/poster/44448",
      "year": 2025,
      "venue": "ICML 2025",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=gscscNNiPN",
      "citations": null,
      "categories": [],
      "id": "gscscNNiPN",
      "track": "main",
      "status": "Poster",
      "keywords": "imbalanced data;consistency;margin bounds;learning theory",
      "tldr": "",
      "primary_area": "general_machine_learning->supervised_learning",
      "similarity_score": 7.223685906676644,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.223685906676644,
      "combined_score": 0.0,
      "rank": 116
    },
    {
      "title": "Modeling SRP-LSH Performance: A Theoretical Framework for Optimizing Approximate Nearest Neighbor Search",
      "authors": [],
      "abstract": "Approximate nearest neighbor (ANN) search in high-dimensional spaces with sign-random-projection locality-sensitive hashing (SRP-LSH) remains challeng\u0002ing due to the lack of principled approaches for configuring its key parameters. We present a theoretical framework that rigorously models SRP-LSH performance and enables principled parameter configuration. At its core is, to our knowledge, the first analytical model that links the number of hash functions and the Hamming distance threshold to search recall, rooted in the binomial distribution of bit colli\u0002sions and the angular similarity distribution of vectors. Building upon this model, we develop an adaptive optimization algorithm that minimizes the candidate set size while satisfying user-specified recall targets. Extensive experiments show that\nour model typically predicts recall with a mean absolute percentage error (MAPE) below 5%. Moreover, our algorithm consistently meets the specified recall targets and simultaneously captures global selectivity trend. Overall, this framework pro\u0002vides a theoretically grounded and practical solution for configuring SRP-LSH in real-world retrieval systems.",
      "url": "https://openreview.net/forum?id=h4hIuid0HY",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "h4hIuid0HY",
      "track": "main",
      "status": "Active",
      "keywords": "Approximate Nearest Neighbor Search; Locality-Sensitive Hashing; Theoretical Analysis; Parameter Optimization; High-Dimensional Retrieval",
      "tldr": "",
      "primary_area": "learning theory",
      "similarity_score": 7.21128816340723,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.21128816340723,
      "combined_score": 0.0,
      "rank": 117
    },
    {
      "title": "Theoretical Study on Multi-objective Heuristic Search",
      "authors": [
        "Shawn Skyler",
        "Shahaf Shperberg",
        "Dor Atzmon",
        "Ariel Felner",
        "Oren Salzman",
        "Shao-Hung Chan",
        "Han Zhang",
        "Sven Koenig",
        "William Yeoh",
        "Carlos Hernandez Ulloa"
      ],
      "abstract": "This paper provides a theoretical study on Multi-Objective Heuristic Search. We first classify states in the state space into must-expand, maybe-expand, and never-expand states and then transfer these definitions to nodes in the search tree. We then formalize a framework that generalizes A* to Multi-Objective Search. We study different ways to order nodes under this framework and their relation to traditional tie-breaking policies and provide theoretical findings. Finally, we study and empirically compare different ordering functions.",
      "url": "https://www.ijcai.org/proceedings/2024/776",
      "year": 2024,
      "venue": "IJCAI 2024",
      "source": "offline_ijcai",
      "doi": null,
      "pdf_url": "https://www.ijcai.org/proceedings/2024/0776.pdf",
      "citations": null,
      "categories": [],
      "id": "paper776",
      "track": "main",
      "status": "Poster",
      "keywords": "Search: S: Heuristic search; Search: S: Other; Search: General",
      "tldr": "",
      "primary_area": "Search",
      "similarity_score": 7.177733969326617,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.177733969326617,
      "combined_score": 0.0,
      "rank": 118
    },
    {
      "title": "A Unified Theoretical Framework for Understanding Difficult-to-learn Examples in Contrastive Learning",
      "authors": [
        "Yi-Ge Zhang",
        "Jingyi Cui",
        "Qiran Li",
        "Yisen Wang"
      ],
      "abstract": "Unsupervised contrastive learning has shown significant performance improvements in recent years, often approaching or even rivaling supervised learning in various tasks. However, its learning mechanism is fundamentally different from that of supervised learning. Previous works have shown that difficult-to-learn examples (well-recognized in supervised learning as examples around the decision boundary), which are essential in supervised learning, contribute minimally in unsupervised settings. In this paper, perhaps surprisingly, we find that the direct removal of difficult-to-learn examples, although reduces the sample size, can boost the downstream classification performance of contrastive learning. To uncover the reasons behind this, we develop a theoretical framework modeling the similarity between different pairs of samples. Guided by this theoretical framework, we conduct a thorough theoretical analysis revealing that the presence of difficult-to-learn examples negatively affects the generalization of contrastive learning. Furthermore, we demonstrate that the removal of these examples, and techniques such as margin tuning and temperature scaling can enhance its generalization bounds, thereby improving performance.\nEmpirically, we propose a simple and efficient mechanism for selecting difficult-to-learn examples and validate the effectiveness of the aforementioned methods, which substantiates the reliability of our proposed theoretical framework.",
      "url": "https://openreview.net/forum?id=P4WnvhVmPV",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "P4WnvhVmPV",
      "track": "main",
      "status": "Reject",
      "keywords": "Machine Learning; Contrastive Learning; Difficult-to-learn Examples",
      "tldr": "",
      "primary_area": "unsupervised, self-supervised, semi-supervised, and supervised representation learning",
      "similarity_score": 7.129651427706412,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.129651427706412,
      "combined_score": 0.0,
      "rank": 119
    },
    {
      "title": "Bernoulli-LoRA: A Theoretical Framework for Randomized Low-Rank Adaptation",
      "authors": [],
      "abstract": "Parameter-efficient fine-tuning (PEFT) has emerged as a crucial approach for adapting large foundational models to specific tasks, particularly as model sizes continue to grow exponentially. Among PEFT methods, Low-Rank Adaptation (LoRA) [Hu et al., 2021] stands out for its effectiveness and simplicity, expressing adaptations as a product of two low-rank matrices. While extensive empirical studies demonstrate LoRA's practical utility, the theoretical understanding of such methods remains limited. Recent work on RAC-LoRA [Malinovsky et al., 2024] took initial steps toward rigorous analysis. In this work, we introduce Bernoulli-LoRA, a novel theoretical framework that unifies and extends existing LoRA approaches. Our method introduces a probabilistic Bernoulli mechanism for selecting which matrix to update. This approach encompasses and generalizes various existing update strategies while maintaining theoretical tractability. Under standard assumptions from non-convex optimization literature, we analyze several variants of our framework: Bernoulli-LoRA-GD, Bernoulli-LoRA-SGD, Bernoulli-LoRA-PAGE, Bernoulli-LoRA-MVR, Bernoulli-LoRA-QGD, Bernoulli-LoRA-MARINA, and Bernoulli-LoRA-EF21, establishing convergence guarantees for each variant. Additionally, we extend our analysis to convex non-smooth functions, providing convergence rates for both constant and adaptive (Polyak-type) stepsizes. Through extensive experiments on various tasks, we validate our theoretical findings and demonstrate the practical efficacy of our approach. This work is a step toward developing theoretically grounded yet practically effective PEFT methods.",
      "url": "https://openreview.net/forum?id=oztUriaGPk",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "oztUriaGPk",
      "track": "main",
      "status": "Active",
      "keywords": "Parameter-Efficient Fine-Tuning;Low-Rank Adaptation;Non-convex Optimization;Non-smooth Optimization;Stochastic Optimization;Variance Reduction;Adaptive Stepsizes",
      "tldr": "",
      "primary_area": "optimization",
      "similarity_score": 7.071864279611812,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.071864279611812,
      "combined_score": 0.0,
      "rank": 120
    },
    {
      "title": "Debiasing with Sufficient Projection: A General Theoretical Framework for Vector Representations",
      "authors": [
        "Enze Shi",
        "Lei Ding",
        "Linglong Kong",
        "Bei Jiang"
      ],
      "abstract": "Pre-trained vector representations in natural language processing often inadvertently encode undesirable social biases. Identifying and removing unwanted biased information from vector representation is an evolving and significant challenge. Our study uniquely addresses this issue from the perspective of statistical independence, proposing a framework for reducing bias by transforming vector representations to an unbiased subspace using sufficient projection. The key to our framework lies in its generality: it adeptly mitigates bias across both debiasing and fairness tasks, and across various vector representation types, including word embeddings and output representations of transformer models. Importantly, we establish the connection between debiasing and fairness, offering theoretical guarantees and elucidating our algorithm’s efficacy. Through extensive evaluation of intrinsic and extrinsic metrics, our method achieves superior performance in bias reduction while maintaining high task performance, and offers superior computational efficiency.",
      "url": "https://aclanthology.org/2024.naacl-long.332/",
      "year": 2024,
      "venue": "NAACL 2024",
      "source": "offline_naacl",
      "doi": null,
      "pdf_url": "https://aclanthology.org/2024.naacl-long.332.pdf",
      "citations": null,
      "categories": [],
      "id": "2024.naacl-long.332",
      "track": "main",
      "status": "Long",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 7.0652680539253065,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.0652680539253065,
      "combined_score": 0.0,
      "rank": 121
    },
    {
      "title": "A Theoretical Framework for Rate-Distortion Limits in Learned Image Compression",
      "authors": [],
      "abstract": "We present a novel systematic theoretical framework to analyze the rate-distortion (R-D) limits of learned image compression. While recent neural codecs have achieved remarkable empirical results, their distance from the information-theoretic limit remains unclear. Our work addresses this gap by decomposing the R-D performance loss into three key components: variance estimation, quantization strategy, and context modeling. First, we derive the optimal latent variance as the second moment under a Gaussian assumption, providing a principled alternative to hyperprior-based estimation. Second, we quantify the gap between uniform quantization and the Gaussian test channel derived from the reverse water-filling theorem. Third, we extend our framework to include context modeling, and demonstrate that accurate mean prediction yields substantial entropy reduction. Unlike prior R-D estimators, our method provides a structurally interpretable perspective that aligns with real compression modules and enables fine-grained analysis. Through joint simulation and end-to-end training, we derive a tight and actionable approximation of the theoretical R-D limits, offering new insights into the design of more efficient learned compression systems.",
      "url": "https://openreview.net/forum?id=YzHbFwYmE1",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "YzHbFwYmE1",
      "track": "main",
      "status": "Active",
      "keywords": "Rate-Distortion Limits;Learned Image Compression;Information Theory;Reverse Water-filling;Context Modeling",
      "tldr": "",
      "primary_area": "learning theory",
      "similarity_score": 7.033795719682257,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 7.033795719682257,
      "combined_score": 0.0,
      "rank": 122
    },
    {
      "title": "A Unified Framework for Fair Graph Generation: Theoretical Guarantees and Empirical Advances",
      "authors": [
        "Zichong Wang",
        "Zhipeng Yin",
        "Wenbin Zhang"
      ],
      "abstract": "Graph generation models play pivotal roles in many real-world applications, from data augmentation to privacy-preserving. Despite their deployment successes, existing approaches often exhibit fairness issues, limiting their adoption in high-risk decision-making applications. Most existing fair graph generation works are based on autoregressive models that suffer from ordering sensitivity, while primarily addressing structural bias and overlooking the critical issue of feature bias. To this end, we propose FairGEM, a novel one-shot graph generation framework designed to mitigate both graph structural bias and node feature bias simultaneously. Furthermore, our theoretical analysis establishes that FairGEM delivers substantially stronger fairness guarantees than existing models while preserving generation quality. Extensive experiments across multiple real-world datasets demonstrate that FairGEM achieves superior performance in both generation quality and fairness.",
      "url": "https://openreview.net/forum?id=T85ADT8a2y",
      "year": 2025,
      "venue": "NIPS 2025",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "T85ADT8a2y",
      "track": "main",
      "status": "Poster",
      "keywords": "Fairness;Graph Generation;GNN",
      "tldr": "",
      "primary_area": "social_and_economic_aspects_of_machine_learning",
      "similarity_score": 6.984273848673643,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.984273848673643,
      "combined_score": 0.0,
      "rank": 123
    },
    {
      "title": "On the Limits of Sparse Autoencoders: A Theoretical Framework and Reweighted Remedy",
      "authors": [],
      "abstract": "Sparse autoencoders (SAEs) have recently emerged as a powerful tool for interpreting the features learned by large language models (LLMs). By reconstructing features with sparsely activated networks, SAEs aim to recover complex superposed polysemantic features into interpretable monosemantic ones. Despite their wide applications, it remains unclear under what conditions SAEs can fully recover the ground truth monosemantic features from the superposed polysemantic ones. In this paper, we provide the first theoretical analysis with a closed-form solution for SAEs, revealing that they generally fail to fully recover the ground truth monosemantic features unless the ground truth features are extremely sparse. To improve the feature recovery of SAEs in general cases, we propose a reweighting strategy targeting at enhancing the reconstruction of the ground truth monosemantic features instead of the observed polysemantic ones. We further establish a theoretical weight selection principle for our proposed weighted SAE (WSAE). Experiments across multiple settings validate our theoretical findings and demonstrate that our WSAE significantly improves feature monosemanticity and interpretability.",
      "url": "https://openreview.net/forum?id=DSOTgzeH3w",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "DSOTgzeH3w",
      "track": "main",
      "status": "Active",
      "keywords": "sparse autoencoder;SAE;theoretical understanding",
      "tldr": "",
      "primary_area": "interpretability and explainable AI",
      "similarity_score": 6.944838513971509,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.944838513971509,
      "combined_score": 0.0,
      "rank": 124
    },
    {
      "title": "Bridging Debiasing Tasks with Sufficient Projection: A General Theoretical Framework for Vector Representations",
      "authors": [
        "Enze Shi",
        "Lei Ding",
        "Linglong Kong",
        "Bei Jiang"
      ],
      "abstract": "Pre-trained vector representations in natural language processing often inadvertently encode undesirable social biases. Identifying and removing unwanted biased information from vector representation is an evolving and significant challenge. Our study uniquely addresses this issue from the perspective of statistical independence, proposing a framework for reducing bias by transforming vector representations to an unbiased subspace using sufficient projection. The key to our framework lies in its generality: it adeptly mitigates bias across both debiasing and fairness tasks, and across various vector representation types, including word embeddings and output representations of transformer models. Importantly, we establish the connection between debiasing and fairness, offering theoretical guarantees and elucidating our algorithm's efficacy. Through extensive evaluation of intrinsic and extrinsic metrics, our method achieves superior performance in bias reduction while maintaining high task performance, and offers superior computational efficiency.",
      "url": "https://openreview.net/forum?id=n3ZXEQKRbO",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "n3ZXEQKRbO",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Gender Debias; Vector Representation; NLP; Algorithmic Fairness",
      "tldr": "",
      "primary_area": "societal considerations including fairness, safety, privacy",
      "similarity_score": 6.927706413393027,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.927706413393027,
      "combined_score": 0.0,
      "rank": 125
    },
    {
      "title": "A Theoretical Framework for Escaping Local Optima in MSE toward Global Convergence",
      "authors": [],
      "abstract": "Deep learning models are trained by minimizing loss functions such as mean squared error (MSE) or cross-entropy, but these objectives are highly non-convex. As a result, optimization often encounters local optima, saddle points, or sharp valleys that hinder convergence and generalization. Although many heuristic approaches, such as momentum, Adam, help mitigate these issues, they provide limited theoretical understanding.\nIn this work, we present a theoretical study of the optimization of MSE. We first provide a mathematical characterization of local optima under MSE and contrast them with those of cross-entropy, identifying when and how they arise. Building on this analysis, we introduce a modified optimization algorithm that explicitly accounts for these properties. Unlike heuristic methods, our approach offers theoretical guarantees for avoiding spurious local traps.\nOur experiments show that the proposed method reliably avoids local optima and converges more effectively than existing optimizers in MNIST, CIFAR10 and CIFAR100 with simple CNN. Our work provides both new insight into MSE optimization for training deep networks.",
      "url": "https://openreview.net/forum?id=a8uipkMIZN",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "a8uipkMIZN",
      "track": "main",
      "status": "Active",
      "keywords": "Mean Square Error;Local Optimum;Linear Algebra",
      "tldr": "",
      "primary_area": "optimization",
      "similarity_score": 6.92652515506082,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.92652515506082,
      "combined_score": 0.0,
      "rank": 126
    },
    {
      "title": "Zero-shot Denoising via Neural Compression: Theoretical and algorithmic framework",
      "authors": [
        "Ali Zafari",
        "Xi Chen",
        "Shirin Jalali"
      ],
      "abstract": "Zero-shot denoising aims to denoise observations without access to training samples or clean reference images. This setting is particularly relevant in practical imaging scenarios involving specialized domains such as  medical imaging or biology. In this work, we propose the *Zero-Shot Neural Compression Denoiser* (ZS-NCD), a novel denoising framework based on neural compression. ZS-NCD treats a neural compression network as an untrained model, optimized directly on patches extracted from a single noisy image. The final reconstruction is then obtained by aggregating the outputs of the trained model over overlapping patches. Thanks to the built-in entropy constraints of compression architectures, our method naturally avoids overfitting and does not require manual regularization or early stopping. Through extensive experiments, we show that ZS-NCD achieves state-of-the-art performance among zero-shot denoisers for both Gaussian and Poisson noise, and generalizes well to both natural and non-natural images. Additionally, we provide new finite-sample theoretical results that characterize upper bounds on the achievable reconstruction error of general maximum-likelihood compression-based denoisers. These results further establish the theoretical foundations of compression-based denoising. Our code is available at: https://github.com/Computational-Imaging-RU/ZS-NCDenoiser.",
      "url": "https://openreview.net/forum?id=DwZD97uHgm",
      "year": 2025,
      "venue": "NIPS 2025",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "DwZD97uHgm",
      "track": "main",
      "status": "Spotlight",
      "keywords": "compression-based denoising;zero-shot image denoising;neural compression",
      "tldr": "",
      "primary_area": "general_machine_learning",
      "similarity_score": 6.886680553150342,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.886680553150342,
      "combined_score": 0.0,
      "rank": 127
    },
    {
      "title": "Reproducibility in Optimization: Theoretical Framework and Limits",
      "authors": [
        "Kwangjun Ahn",
        "Prateek Jain",
        "Ziwei Ji",
        "Satyen Kale",
        "Praneeth Netrapalli",
        "Gil I. Shamir"
      ],
      "abstract": " We initiate a formal study of reproducibility in optimization. We define a quantitative measure of reproducibility of optimization procedures in the face of noisy or error-prone operations such as inexact or stochastic gradient computations or inexact initialization. We then analyze several convex optimization settings of interest such as smooth, non-smooth, and strongly-convex objective functions and establish tight bounds on the limits of reproducibility in each setting. Our analysis reveals a fundamental trade-off between computation and reproducibility: more computation is necessary (and sufficient) for better reproducibility.",
      "url": "https://nips.cc/virtual/2022/poster/54471",
      "year": 2022,
      "venue": "NIPS 2022",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=3LMI8CHDb0g",
      "citations": null,
      "categories": [],
      "id": "3LMI8CHDb0g",
      "track": "main",
      "status": "Accept",
      "keywords": "reproducibility;first-order optimization;convex optimization;inexact gradient oracles",
      "tldr": "We initiate a formal study of reproducibility in optimization by defining a quantitative measure and characterizing the fundamental limits for various settings.",
      "primary_area": "",
      "similarity_score": 6.874599208394791,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.874599208394791,
      "combined_score": 0.0,
      "rank": 128
    },
    {
      "title": "RAC-LoRA: A Theoretical Optimization Framework for Low-Rank Adaptation",
      "authors": [
        "Grigory Malinovsky",
        "Umberto Michieli",
        "Hasan Abed Al Kader Hammoud",
        "Taha Ceritli",
        "Hayder Elesedy",
        "Peter Richtárik",
        "Mete Ozay"
      ],
      "abstract": "Fine-tuning has become a popular approach to adapting large foundational models to specific tasks. As the size of models and datasets grows, parameter-efficient fine-tuning techniques are increasingly important. One of the most widely used methods is Low-Rank Adaptation (LoRA), with adaptation update expressed as the product of two low-rank matrices. While LoRA was shown to possess strong performance in fine-tuning, it often underperforms when compared to full-parameter fine-tuning (FPFT). Although many variants of LoRA have been extensively studied empirically, their theoretical optimization analysis is heavily under-explored. The starting point of our work is a demonstration that LoRA and its two extensions, Asymmetric LoRA and Chain of LoRA, indeed encounter convergence issues. To address these issues, we propose a general optimization framework that rigorously analyzes the convergence rates of LoRA-based methods. Our approach inherits the empirical benefits of LoRA-style heuristics, but introduces several small but important algorithmic modifications which turn it into a provably convergent method. Our framework serves as a bridge between FPFT and low-rank adaptation. We provide provable guarantees of convergence to the same solution as FPFT, along with the rate of convergence. Additionally, we present a convergence analysis for smooth, non-convex loss functions, covering gradient descent, stochastic gradient descent, and federated learning settings. Our theoretical findings are supported by experimental results.",
      "url": "https://openreview.net/forum?id=VSKV3GykuE",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "VSKV3GykuE",
      "track": "main",
      "status": "Reject",
      "keywords": "LORA;optimization;stochastic optimization;low-rank adaptation",
      "tldr": "",
      "primary_area": "optimization",
      "similarity_score": 6.847574430479554,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.847574430479554,
      "combined_score": 0.0,
      "rank": 129
    },
    {
      "title": "Fault-Tolerant Federated Reinforcement Learning with Theoretical Guarantee",
      "authors": [
        "Flint Xiaofeng Fan",
        "Yining Ma",
        "Zhongxiang Dai",
        "Wei Jing",
        "Cheston Tan",
        "Bryan Kian Hsiang Low"
      ],
      "abstract": "The growing literature of Federated Learning (FL) has recently inspired Federated Reinforcement Learning (FRL) to encourage multiple agents to federatively build a better decision-making policy without sharing raw trajectories. Despite its promising applications, existing works on FRL fail to I) provide theoretical analysis on its convergence, and II) account for random system failures and adversarial attacks. Towards this end, we propose the first FRL framework the convergence of which is guaranteed and tolerant to less than half of the participating agents being random system failures or adversarial attackers. We prove that the sample efficiency of the proposed framework is guaranteed to improve with the number of agents and is able to account for such potential failures or attacks. All theoretical results are empirically verified on various RL benchmark tasks.",
      "url": "https://nips.cc/virtual/2021/poster/28144",
      "year": 2021,
      "venue": "NIPS 2021",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=ospGnpuf6L",
      "citations": null,
      "categories": [],
      "id": "ospGnpuf6L",
      "track": "main",
      "status": "Poster",
      "keywords": "Reinforcement learning;federated learning;Byzantine-tolerant optimization",
      "tldr": "This paper provides the theoretical ground to study the sample efficiency of Federated Reinforcement Learning with respect to the number of participating agents, accounting for faulty agents.",
      "primary_area": "",
      "similarity_score": 6.807989961927655,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.807989961927655,
      "combined_score": 0.0,
      "rank": 130
    },
    {
      "title": "A Theoretical Framework for Inference Learning",
      "authors": [
        "Nicholas Alonso",
        "Beren Millidge",
        "Jeffrey Krichmar",
        "Emre Neftci"
      ],
      "abstract": "Backpropagation (BP) is the most successful and widely used algorithm in deep learning. However, the computations required by BP are challenging to reconcile with known neurobiology. This difficulty has stimulated interest in more biologically plausible alternatives to BP. One such algorithm is the inference learning algorithm (IL). IL trains predictive coding models of neural circuits and has achieved equal performance to BP on supervised and auto-associative tasks. In contrast to BP, however, the mathematical foundations of IL are not well-understood. Here, we develop a novel theoretical framework for IL. Our main result is that IL closely approximates an optimization method known as implicit stochastic gradient descent (implicit SGD), which is distinct from the explicit SGD implemented by BP. Our results further show how the standard implementation of IL can be altered to better approximate implicit SGD. Our novel implementation considerably improves the stability of IL across learning rates, which is consistent with our theory, as a key property of implicit SGD is its stability. We provide extensive simulation results that further support our theoretical interpretations and find IL achieves quicker convergence when trained with mini-batch size one while performing competitively with BP for larger mini-batches when combined with Adam.",
      "url": "https://nips.cc/virtual/2022/poster/53058",
      "year": 2022,
      "venue": "NIPS 2022",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=7yJMZwhIC2k",
      "citations": null,
      "categories": [],
      "id": "7yJMZwhIC2k",
      "track": "main",
      "status": "Accept",
      "keywords": "Predictive Coding;Backpropagation;Synaptic Plasticity;Local Learning;Inference Learning",
      "tldr": "In this paper, we develop a novel theoretical framework for inference learning, a biologically plausible local learning algorithm for deep neural networks.",
      "primary_area": "",
      "similarity_score": 6.807427371677286,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.807427371677286,
      "combined_score": 0.0,
      "rank": 131
    },
    {
      "title": "Functional Regularization for Representation Learning: A Unified Theoretical Perspective",
      "authors": [
        "Siddhant Garg",
        "Yingyu Liang"
      ],
      "abstract": "Unsupervised and self-supervised learning approaches have become a crucial tool to learn representations for downstream prediction tasks. While these approaches are widely used in practice and achieve impressive empirical gains, their theoretical understanding largely lags behind. Towards bridging this gap, we present a unifying perspective where several such approaches can be viewed as imposing a regularization on the representation via a learnable function using unlabeled data. We propose a discriminative theoretical framework for analyzing the sample complexity of these approaches, which generalizes the framework of (Balcan and Blum, 2010) to allow learnable regularization functions. Our sample complexity bounds show that, with carefully chosen hypothesis classes to exploit the structure in the data, these learnable regularization functions can prune the hypothesis space, and help reduce the amount of labeled data needed. We then provide two concrete examples of functional regularization, one using auto-encoders and the other using masked self-supervision, and apply our framework to quantify the reduction in the sample complexity bound of labeled data. We also provide complementary empirical results to support our analysis.",
      "url": "https://nips.cc/virtual/2020/poster/17555",
      "year": 2020,
      "venue": "NIPS 2020",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2020/file/c793b3be8f18731f2a4c627fb3c6c63d-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "17555",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 6.785341974588365,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.785341974588365,
      "combined_score": 0.0,
      "rank": 132
    },
    {
      "title": "Connecting Data to Mechanisms with Meta Structual Causal Model",
      "authors": [
        "Gong Heyang"
      ],
      "abstract": "Recent years have seen impressive progress in theoretical and algorithmic developments of causal inference across various disciplines in science and engineering. However, there is still some unresolved theoretical problems, especially for cyclic causal relationships. In this article, we propose a meta structure causal model (Meta-SCM) framework inspired by understanding causality as information transfer. A key feature of our framework is the introduction of the concept of \\emph{active mechanisms} to connect data and the collection of underlying causal mechanisms. We show that the Meta-SCM provides a novel approach to address the theoretical complications for modeling cyclic causal relations. In addition, we propose a \\emph{sufficient activated mechanisms} assumption, and explain its relationship with existing assumptions in causal representation learning.  Finally, we conclude the main idea of the meta-SCM framework with an emphasis on its theoretical and conceptual novelty.",
      "url": "https://openreview.net/forum?id=gggnCQBT_iE",
      "year": 2022,
      "venue": "ICLR 2022",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "gggnCQBT_iE",
      "track": "main",
      "status": "Reject",
      "keywords": "meta-SCM;cyclic causal models;sufficient activated mechanisms",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 6.729915685380014,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.729915685380014,
      "combined_score": 0.0,
      "rank": 133
    },
    {
      "title": "polybasic Speculative Decoding Through a Theoretical Perspective",
      "authors": [
        "Ruilin Wang",
        "Huixia Li",
        "Yuexiao Ma",
        "Xiawu Zheng",
        "Fei Chao",
        "Xuefeng Xiao",
        "Rongrong Ji"
      ],
      "abstract": "Inference latency stands as a critical bottleneck in the large-scale deployment of Large Language Models (LLMs). Speculative decoding methods have recently shown promise in accelerating inference without compromising the output distribution. However, existing work typically relies on a dualistic draft-verify framework and lacks rigorous theoretical grounding. In this paper, we introduce a novel \\emph{polybasic} speculative decoding framework, underpinned by a comprehensive theoretical analysis. Specifically, we prove a fundamental theorem that characterizes the optimal inference time for multi-model speculative decoding systems, shedding light on how to extend beyond the dualistic approach to a more general polybasic paradigm. Through our theoretical investigation of multi-model token generation, we expose and optimize the interplay between model capabilities, acceptance lengths, and overall computational cost. Our framework supports both standalone implementation and integration with existing speculative techniques, leading to accelerated performance in practice. Experimental results across multiple model families demonstrate that our approach yields speedup ratios ranging from $3.31\\times$ to $4.01\\times$ for LLaMA2-Chat 7B, up to $3.87 \\times$ for LLaMA3-8B, up to $4.43 \\times$ for Vicuna-7B and up to $3.85 \\times$ for Qwen2-7B---all while preserving the original output distribution. We release our theoretical proofs and implementation code to facilitate further investigation into polybasic speculative decoding.",
      "url": "https://icml.cc/virtual/2025/poster/45669",
      "year": 2025,
      "venue": "ICML 2025",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=JrxJUMqqz4",
      "citations": null,
      "categories": [],
      "id": "JrxJUMqqz4",
      "track": "main",
      "status": "Poster",
      "keywords": "speculative decoding",
      "tldr": "",
      "primary_area": "deep_learning->large_language_models",
      "similarity_score": 6.697254435109499,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.697254435109499,
      "combined_score": 0.0,
      "rank": 134
    },
    {
      "title": "Provable Reward-Agnostic Preference-Based Reinforcement Learning",
      "authors": [
        "Wenhao Zhan",
        "Masatoshi Uehara",
        "Wen Sun",
        "Jason D. Lee"
      ],
      "abstract": "Preference-based Reinforcement Learning (PbRL) is a paradigm in which an RL agent learns to optimize a task using pair-wise preference-based feedback over trajectories, rather than explicit reward signals. While PbRL has demonstrated practical success in fine-tuning language models, existing theoretical work focuses on regret minimization and fails to capture most of the practical frameworks. In this study, we fill in such a gap between theoretical PbRL and practical algorithms by proposing a theoretical reward-agnostic PbRL framework where exploratory trajectories that enable accurate learning of hidden reward functions are acquired before collecting any human feedback. Theoretical analysis demonstrates that our algorithm requires less human feedback for learning the optimal policy under preference-based models with linear parameterization and unknown transitions, compared to the existing theoretical literature. Specifically, our framework can incorporate linear and low-rank MDPs with efficient sample complexity. Additionally, we investigate reward-agnostic RL with action-based comparison feedback and introduce an efficient querying algorithm tailored to this scenario.",
      "url": "https://iclr.cc/virtual/2024/poster/17417",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=yTBXeXdbMf",
      "citations": null,
      "categories": [],
      "id": "yTBXeXdbMf",
      "track": "main",
      "status": "Spotlight",
      "keywords": "reinforcement learning theory;reward-agnostic learning",
      "tldr": "",
      "primary_area": "reinforcement learning",
      "similarity_score": 6.652836725184391,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.652836725184391,
      "combined_score": 0.0,
      "rank": 135
    },
    {
      "title": "Theoretical Analysis of Contrastive Learning under Imbalanced Data: From Training Dynamics to a Pruning Solution",
      "authors": [],
      "abstract": "Contrastive learning has emerged as a powerful framework for learning generalizable representations, yet its theoretical understanding remains limited, particularly under imbalanced data distributions that are prevalent in real-world applications. Such an imbalance can degrade representation quality and induce biased model behavior, yet a rigorous characterization of these effects is lacking. In this work, we develop a theoretical framework to analyze the training dynamics of contrastive learning with Transformer-based encoders under imbalanced data. Our results reveal that neuron weights evolve through three distinct stages of training, with different dynamics for majority features, minority features, and noise. We further show that minority features reduce representational capacity, increase the need for more complex architectures, and hinder the separation of ground-truth features from noise. Inspired by these neuron-level behaviors, we show that pruning restores performance degraded by imbalance and enhances feature separation, offering both conceptual insights and practical guidance. Major theoretical findings are validated through numerical experiments.",
      "url": "https://openreview.net/forum?id=DUXG9E8dEO",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "DUXG9E8dEO",
      "track": "main",
      "status": "Active",
      "keywords": "Contrastive learning;Feature learning;Training dynamics;Theoretical analysis",
      "tldr": "",
      "primary_area": "learning theory",
      "similarity_score": 6.643958462535805,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.643958462535805,
      "combined_score": 0.0,
      "rank": 136
    },
    {
      "title": "Theoretical Modeling of Large Language Model Self-Improvement Training Dynamics Through Solver-Verifier Gap",
      "authors": [],
      "abstract": "Self-improvement is a significant techniques within the realm of large language model (LLM), aiming to enhance the LLM performance without relying on external data. Despite its significance, generally how LLM performances evolve during the self-improvement process remains underexplored. In this paper, we theoretically model the training dynamics of self-improvement via the concept of solver-verifier gap. This is inspired by the conjecture that the performance enhancement of self-improvement stems from the gap between LLM's solver capability and verifier capability. Based on the theoretical framework, we further show how to model the entire training trajectory. This framework allows quantifying the capability limit of self-improvement by fitting the theoretical model to the experiment results. We validate the effectiveness of the theoretical framework on various LLMs and datasets. Beyond self-improvement, we extend our analysis to investigate how external data influences these dynamics within the framework. Notably, we find that under limited external data regimes, such external data can be utilized at any stage without significantly affecting final performances, which accords with the empirical observations.",
      "url": "https://openreview.net/forum?id=Hh7x3c0cZl",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "Hh7x3c0cZl",
      "track": "main",
      "status": "Active",
      "keywords": "Training Dynamics;Self-Improvement",
      "tldr": "",
      "primary_area": "foundation or frontier models, including LLMs",
      "similarity_score": 6.631310694839554,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.631310694839554,
      "combined_score": 0.0,
      "rank": 137
    },
    {
      "title": "Probabilistic Inference with Algebraic Constraints: Theoretical Limits and Practical Approximations",
      "authors": [
        "Zhe Zeng",
        "Paolo Morettin",
        "Fanqi Yan",
        "Antonio Vergari",
        "Guy Van den Broeck"
      ],
      "abstract": "Weighted model integration (WMI) is a framework to perform advanced probabilistic inference on hybrid domains, i.e., on distributions over mixed continuous-discrete random variables and in presence of complex logical and arithmetic constraints. In this work, we advance the WMI framework on both the theoretical and algorithmic side. First, we exactly trace the boundaries of tractability for WMI inference by proving that to be amenable to exact and efficient inference a WMI problem has to posses a tree-shaped structure with logarithmic diameter. While this result deepens our theoretical understanding of WMI it hinders the practical applicability of exact WMI solvers to real-world problems. To overcome this, we propose the first approximate WMI solver that does not resort to sampling, but performs exact inference on one approximate models. Our solution performs message passing in a relaxed problem structure iteratively to recover certain lost dependencies and, as our experiments suggest, is competitive with other SOTA WMI solvers.",
      "url": "https://nips.cc/virtual/2020/poster/18846",
      "year": 2020,
      "venue": "NIPS 2020",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2020/file/85934679f30131d812a8c7475a7d0f74-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "18846",
      "track": "main",
      "status": "Spotlight",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 6.617036264041408,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.617036264041408,
      "combined_score": 0.0,
      "rank": 138
    },
    {
      "title": "Connecting NTK and NNGP: A Unified Theoretical Framework for Neural Network Learning Dynamics in the Kernel Regime",
      "authors": [
        "Yehonatan Avidan",
        "Qianyi Li",
        "Haim Sompolinsky"
      ],
      "abstract": "Artificial neural networks (ANNs) have revolutionized machine learning in recent years, but a complete theoretical framework for their learning process is still lacking. Substantial theoretical advances have been achieved for infinitely wide networks. In this regime, two disparate theoretical frameworks have been used, in which the network’s output is described using kernels: one framework is based on the Neural Tangent Kernel (NTK), which assumes linearized gradient descent dynamics, while the Neural Network Gaussian Process (NNGP) kernel assumes a Bayesian framework. However, the relation between these two frameworks and between their underlying sets of assumptions has remained elusive. This work unifies these two distinct theories using gradient descent learning dynamics with an additional small noise in an ensemble of randomly initialized infinitely wide deep networks. We derive an exact analytical expression for the network input-output function during and after learning and introduce a new time-dependent Neural Dynamical Kernel (NDK) from which both NTK and NNGP kernels can be derived. We identify two important learning phases characterized by different time scales: gradient-driven and diffusive learning. In the initial gradient-driven learning phase, the dynamics is dominated by deterministic gradient descent, and is adequately described by the NTK theory. This phase is followed by the slow diffusive learning stage, during which the network parameters sample the solution space, ultimately approaching the equilibrium posterior distribution corresponding to NNGP. Combined with numerical evaluations on synthetic and benchmark datasets, we provide novel insights into the different roles of initialization, regularization, and network depth, as well as phenomena such as early stopping and representational drift. This work closes the gap between the NTK and NNGP theories, providing a comprehensive framework for understanding the learning process of deep neural networks in the infinite width limit.",
      "url": "https://openreview.net/forum?id=5EtSvYUU0v",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "5EtSvYUU0v",
      "track": "main",
      "status": "Reject",
      "keywords": "Learning dynamics;Neural tangent kernel;Neural network Gaussian process;Infinite width limit;Representational drift;Statistical mechanics",
      "tldr": "",
      "primary_area": "learning theory",
      "similarity_score": 6.602112086525191,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.602112086525191,
      "combined_score": 0.0,
      "rank": 139
    },
    {
      "title": "A Theoretical Framework for Target Propagation",
      "authors": [
        "Alexander Meulemans",
        "Francesco Carzaniga",
        "Johan Suykens",
        "João Sacramento",
        "Benjamin F. Grewe"
      ],
      "abstract": "The success of deep learning, a brain-inspired form of AI, has sparked interest in understanding how the brain could similarly learn across multiple layers of neurons. However, the majority of biologically-plausible learning algorithms have not yet reached the performance of backpropagation (BP), nor are they built on strong theoretical foundations. Here, we analyze target propagation (TP), a popular but not yet fully understood alternative to BP, from the standpoint of mathematical optimization. Our theory shows that TP is closely related to Gauss-Newton optimization and thus substantially differs from BP. Furthermore, our analysis reveals a fundamental limitation of difference target propagation (DTP), a well-known variant of TP, in the realistic scenario of non-invertible neural networks. We provide a first solution to this problem through a novel reconstruction loss that improves feedback weight training, while simultaneously introducing architectural flexibility by allowing for direct feedback connections from the output to each hidden layer. Our theory is corroborated by experimental results that show significant improvements in performance and in the alignment of forward weight updates with loss gradients, compared to DTP.",
      "url": "https://nips.cc/virtual/2020/poster/17648",
      "year": 2020,
      "venue": "NIPS 2020",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2020/file/e7a425c6ece20cbc9056f98699b53c6f-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "17648",
      "track": "main",
      "status": "Spotlight",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 6.568506621086399,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.568506621086399,
      "combined_score": 0.0,
      "rank": 140
    },
    {
      "title": "A Theoretical and Practical Framework for Regression and Classification from Truncated Samples",
      "authors": [
        "Andrew Ilyas",
        "Emmanouil Zampetakis",
        "Constantinos Daskalakis"
      ],
      "abstract": "Machine learning and statistics are invaluable for extracting insights from data. A key assumption of most methods, however, is that they have access to independent samples from the distribution of relevant data. As such, these methods often perform poorly in the face of {\\em biased data} which breaks this assumption. In this work, we consider the classical challenge of bias due to truncation, wherein samples falling outside of an “observation window” cannot be observed. We present a general framework for regression and classification from  samples that are truncated according to the value of the dependent variable. The framework argues that stochastic gradient descent (SGD) can be efficiently executed on the population log-likelihood of the truncated sample. Our framework is  broadly applicable, and we provide end-to-end guarantees for the well-studied problems of truncated logistic and probit regression, where we argue that the true model parameters can be identified computationally and statistically efficiently from truncated data, extending recent work on truncated linear regression. We also provide experiments to illustrate the practicality of our framework on synthetic and real data.",
      "url": "https://proceedings.mlr.press/v108/ilyas20a.html",
      "year": 2020,
      "venue": "AISTATS 2020",
      "source": "offline_aistats",
      "doi": null,
      "pdf_url": "http://proceedings.mlr.press/v108/ilyas20a/ilyas20a.pdf",
      "citations": null,
      "categories": [],
      "id": "498e4b69d0",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 6.5429272938295835,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.5429272938295835,
      "combined_score": 0.0,
      "rank": 141
    },
    {
      "title": "Mirror Learning: A Unifying Framework of Policy Optimisation",
      "authors": [
        "Jakub Grudzien",
        "Christian A Schroeder De Witt",
        "Jakob Foerster"
      ],
      "abstract": "Modern deep reinforcement learning (RL) algorithms are motivated by either the general policy improvement (GPI) or trust-region learning (TRL) frameworks. However, algorithms that strictly respect these theoretical frameworks have proven unscalable. Surprisingly, the only known scalable algorithms violate the GPI/TRL assumptions, e.g. due to required regularisation or other heuristics. The current explanation of their empirical success is essentially “by analogy”: they are deemed approximate adaptations of theoretically sound methods. Unfortunately, studies have shown that in practice these algorithms differ greatly from their conceptual ancestors. In contrast, in this paper, we introduce a novel theoretical framework, named Mirror Learning, which provides theoretical guarantees to a large class of algorithms, including TRPO and PPO. While the latter two exploit the flexibility of our framework, GPI and TRL fit in merely as pathologically restrictive corner cases thereof. This suggests that the empirical performance of state-of-the-art methods is a direct consequence of their theoretical properties, rather than of aforementioned approximate analogies. Mirror learning sets us free to boldly explore novel, theoretically sound RL algorithms, a thus far uncharted wonderland.",
      "url": "https://icml.cc/virtual/2022/poster/17441",
      "year": 2022,
      "venue": "ICML 2022",
      "source": "offline_icml",
      "doi": null,
      "pdf_url": "https://proceedings.mlr.press/v162/grudzien22a/grudzien22a.pdf",
      "citations": null,
      "categories": [],
      "id": "17441",
      "track": "main",
      "status": "Spotlight",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 6.5399954498054536,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.5399954498054536,
      "combined_score": 0.0,
      "rank": 142
    },
    {
      "title": "A Generative Model for Game Theory with Flow Equilibrium",
      "authors": [
        "Zhiyu Zhao",
        "David Henry Mguni",
        "Yali Du",
        "Kaiyang Guo",
        "Haifeng Zhang",
        "Jun Wang"
      ],
      "abstract": "In recent years, generative models have emerged as a groundbreaking development in the field of artificial intelligence, transforming various domains such as image synthesis, natural language processing, and data generation. While recent studies have integrated generative models into multi-agent scenarios, their game-theoretical implications have remained largely unexplored. Specifically, the relationship between solutions derived from generative models and game theoretical equilibrium concepts lacks rigorous investigation.\nThis paper aims to bridge the gap between generative models and game theory by introducing a novel probabilistic framework for modelling multi-agent decision-making problems. This innovative framework reinterprets these problems as generative processes. Furthermore, we introduce a training objective known as \"flow equilibrium\" and establish a theoretical connection between flow equilibrium and Nash equilibrium. To analyse the theoretical properties of our framework, we present a tabular version algorithm along with a convergence proof. Additionally, we propose an extended algorithm incorporating neural networks to handle more complex environments. Notably, our framework naturally incorporates opponent modelling. Harnessing the capabilities of generative models, our framework excels in capturing the intricate dynamics of strategic interactions among agents. We validate our approach through testing on various multi-agent tasks, including cooperative and general-sum games. The empirical results consistently support our theoretical findings, demonstrating that our framework consistently outperforms existing methods in terms of solution quality.",
      "url": "https://openreview.net/forum?id=eVlcdbIx2O",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "eVlcdbIx2O",
      "track": "main",
      "status": "Withdraw",
      "keywords": "Generative Model;Variational Inference;Game Theory",
      "tldr": "",
      "primary_area": "general machine learning (i.e., none of the above)",
      "similarity_score": 6.539542305479185,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.539542305479185,
      "combined_score": 0.0,
      "rank": 143
    },
    {
      "title": "Learning on Random Balls is Sufficient for Estimating (Some) Graph Parameters",
      "authors": [
        "Takanori Maehara",
        "Hoang NT"
      ],
      "abstract": "Theoretical analyses for graph learning methods often assume a complete observation of the input graph. Such an assumption might not be useful for handling any-size graphs due to the scalability issues in practice. In this work, we develop a theoretical framework for graph classification problems in the partial observation setting (i.e., subgraph samplings). Equipped with insights from graph limit theory, we propose a new graph classification model that works on a randomly sampled subgraph and a novel topology to characterize the representability of the model. Our theoretical framework contributes a theoretical validation of mini-batch learning on graphs and leads to new learning-theoretic results on generalization bounds as well as size-generalizability without assumptions on the input.",
      "url": "https://nips.cc/virtual/2021/poster/27448",
      "year": 2021,
      "venue": "NIPS 2021",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=Tbq5fYViJzm",
      "citations": null,
      "categories": [],
      "id": "Tbq5fYViJzm",
      "track": "main",
      "status": "Poster",
      "keywords": "graph parameters;Benjamini-Schramm convergence;random sampling;graph learning theory;graph classification;GNN",
      "tldr": "A graph parameter is estimable by GNNs+random sampling if and only if it is continuous in randomized Benjamini-Schramm topology.",
      "primary_area": "",
      "similarity_score": 6.538915079315144,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.538915079315144,
      "combined_score": 0.0,
      "rank": 144
    },
    {
      "title": "A Theoretical Framework for an Efficient Normalizing Flow-Based Solution to the Electronic Schrödinger Equation",
      "authors": [
        "Daniel Freedman",
        "Eyal Rozenberg",
        "Alex Bronstein"
      ],
      "abstract": "A central problem in quantum mechanics involves solving the Electronic Schrödinger Equation for a molecule or material. The Variational Monte Carlo approach to this problem approximates a particular variational objective via sampling, and then optimizes this approximated objective over a chosen parameterized family of wavefunctions, known as the ansatz. Recently neural networks have been used as the ansatz, with accompanying success. However, sampling from such wavefunctions has required the use of a Markov Chain Monte Carlo approach, which is inherently inefficient. In this work, we propose a solution to this problem via an ansatz which is cheap to sample from, yet satisfies the requisite quantum mechanical properties. We prove that a normalizing flow using the following two essential ingredients satisfies our requirements: (a) a base distribution which is constructed from Determinantal Point Processes; (b) flow layers which are equivariant to a particular subgroup of the permutation group. We then show how to construct both continuous and discrete normalizing flows which satisfy the requisite equivariance. We further demonstrate the manner in which the non-smooth nature (``cusps'') of the wavefunction may be captured, and how the framework may be generalized to provide induction across multiple molecules. The resulting theoretical framework entails an efficient approach to solving the Electronic Schrödinger Equation.",
      "url": "https://ojs.aaai.org/index.php/AAAI/article/view/31996",
      "year": 2025,
      "venue": "AAAI 2025",
      "source": "offline_aaai",
      "doi": null,
      "pdf_url": "https://ojs.aaai.org/index.php/AAAI/article/view/31996/34151",
      "citations": null,
      "categories": [],
      "id": "article-31996",
      "track": "main",
      "status": "Technical",
      "keywords": "",
      "tldr": "",
      "primary_area": "application domains",
      "similarity_score": 6.527554716527314,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.527554716527314,
      "combined_score": 0.0,
      "rank": 145
    },
    {
      "title": "SUPER-ADAM: Faster and Universal Framework of Adaptive Gradients",
      "authors": [
        "Feihu Huang",
        "Junyi Li",
        "Heng Huang"
      ],
      "abstract": "Adaptive gradient methods have shown excellent performances for solving many machine learning problems. Although multiple adaptive gradient  methods were recently studied, they mainly focus on either empirical or theoretical aspects and also only work for specific problems by using some  specific adaptive learning rates. Thus, it is desired to design  a  universal  framework for practical algorithms of adaptive gradients with theoretical guarantee to solve general problems. To fill this gap, we propose a faster and universal framework of adaptive gradients (i.e., SUPER-ADAM) by introducing a universal adaptive matrix that includes most existing adaptive gradient forms. Moreover, our framework can flexibly integrate the momentum and variance reduced techniques. In particular, our novel framework provides the convergence analysis support for adaptive gradient methods under the nonconvex setting. In theoretical analysis, we prove that our SUPER-ADAM algorithm can achieve the best known gradient (i.e., stochastic first-order oracle (SFO)) complexity of $\\tilde{O}(\\epsilon^{-3})$ for finding an $\\epsilon$-stationary point of nonconvex optimization, which matches the lower bound for stochastic smooth nonconvex optimization. In numerical experiments, we employ various deep learning tasks to validate that our algorithm consistently outperforms the existing adaptive algorithms. Code is available at https://github.com/LIJUNYI95/SuperAdam",
      "url": "https://nips.cc/virtual/2021/poster/27440",
      "year": 2021,
      "venue": "NIPS 2021",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=nFdJSm9dy83",
      "citations": null,
      "categories": [],
      "id": "nFdJSm9dy83",
      "track": "main",
      "status": "Poster",
      "keywords": "Adaptive Gradient;Adam;Universal Framework;Nonconvex Optimization;Deep Learning",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 6.46498798373444,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 6.46498798373444,
      "combined_score": 0.0,
      "rank": 146
    },
    {
      "title": "Hyper-Connections",
      "authors": [
        "Defa Zhu",
        "Hongzhi Huang",
        "Zihao Huang",
        "Yutao Zeng",
        "Yunyao Mao",
        "Banggu Wu",
        "Qiyang Min",
        "Xun Zhou"
      ],
      "abstract": "We present hyper-connections, a simple yet effective method that can serve as an alternative to residual connections. This approach specifically addresses common drawbacks observed in residual connection variants, such as the seesaw effect between gradient vanishing and representation collapse. Theoretically, hyper-connections allow the network to adjust the strength of connections between features at different depths and dynamically rearrange layers. We conduct experiments focusing on the pre-training of large language models, including dense and sparse models, where hyper-connections show significant performance improvements over residual connections. Additional experiments conducted on vision tasks also demonstrate similar improvements. We anticipate that this method will be broadly applicable and beneficial across a wide range of AI problems.",
      "url": "https://iclr.cc/virtual/2025/poster/30709",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=9FqARW7dwB",
      "citations": null,
      "categories": [],
      "id": "9FqARW7dwB",
      "track": "main",
      "status": "Poster",
      "keywords": "Network Architecture;Residual Connections;LLMs;Pre-training",
      "tldr": "",
      "primary_area": "foundation or frontier models, including LLMs",
      "similarity_score": 0.449487223185653,
      "novelty_score": 0.9807692307692308,
      "recency_score": 0.9,
      "relevance_score": 0.6148461669556959,
      "bm25_score": 1.0,
      "combined_score": 0.6148461669556959,
      "rank": 147
    },
    {
      "title": "Value-Alignment via Safe Semantic Manifold-Constrained Latent Diffusion",
      "authors": [],
      "abstract": "LLM-based detoxification often shifts explicit toxicity into subtler forms: profanities vanish while harm persists through insinuations, stereotypes, microaggressions, and subtly discriminatory framing. We reformulate detoxification from a value-alignment perspective as a multi-principle constrained generation problem that enforces explicit harmlessness, guards against insinuations, stereotypes, microaggressions, and subtly discriminatory framing, and simultaneously maintains fairness consistency, and helpfulness. Specifically, we propose a diffusion-based process-level aligner, SafeManifold-Diffusion, a novel framework that combines conditional latent diffusion with a diffusion-map-based semantic safety manifold to enforce both semantic fidelity and value alignment. Given an offensive input, our model generates a rewritten sentence by simulating a denoising trajectory in latent space, conditioned on the original content embedding. To prevent the trajectory from entering semantically toxic regions, we construct a nonlinear safe semantic manifold from verified non-offensive latent representations using diffusion geometry, and constrain generation via explicit manifold projection at each sampling step. Experiments on four detoxification datasets demonstrate that SafeManifold-Diffusion achieves state-of-the-art performance in reducing both explicit and implicit toxicity while preserving intent, improving fairness, and producing more helpful outputs. Our results suggest that aligning generation with structured semantic constraints is crucial for building trustworthy language systems",
      "url": "https://openreview.net/forum?id=A5AejTTloS",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "A5AejTTloS",
      "track": "main",
      "status": "Active",
      "keywords": "value alignment; diffusion model; Manifold-Constrained",
      "tldr": "",
      "primary_area": "alignment, fairness, safety, privacy, and societal considerations",
      "similarity_score": 0.047005554906147454,
      "novelty_score": 0.9701086956521741,
      "recency_score": 1.0,
      "relevance_score": 0.433306431418567,
      "bm25_score": 0.7306825498224092,
      "combined_score": 0.433306431418567,
      "rank": 148
    },
    {
      "title": "ManiPose: Manifold-Constrained Multi-Hypothesis 3D Human Pose Estimation",
      "authors": [
        "Cédric Rommel",
        "Victor Letzelter",
        "Nermin Samet",
        "Renaud Marlet",
        "Matthieu Cord",
        "Patrick Perez",
        "Eduardo Valle"
      ],
      "abstract": "We propose ManiPose, a manifold-constrained multi-hypothesis model for human-pose 2D-to-3D lifting. We provide theoretical and empirical evidence that, due to the depth ambiguity inherent to monocular 3D human pose estimation, traditional regression models suffer from pose-topology consistency issues, which standard evaluation metrics (MPJPE, P-MPJPE and PCK) fail to assess.  ManiPose addresses depth ambiguity by proposing multiple candidate 3D poses for each 2D input, each with its estimated plausibility. Unlike previous multi-hypothesis approaches, ManiPose forgoes generative models, greatly facilitating its training and usage. By constraining the outputs to lie on the human pose manifold, ManiPose guarantees the consistency of all hypothetical poses, in contrast to previous works. We showcase the performance of ManiPose on real-world datasets, where it outperforms state-of-the-art models in pose consistency by a large margin while being very competitive on the MPJPE metric.",
      "url": "https://neurips.cc/virtual/2024/poster/93050",
      "year": 2024,
      "venue": "NIPS 2024",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=xxY8d4rnSb",
      "citations": null,
      "categories": [],
      "id": "xxY8d4rnSb",
      "track": "main",
      "status": "Poster",
      "keywords": "human pose estimation;depth ambiguity;multiple choice learning",
      "tldr": "",
      "primary_area": "machine_vision",
      "similarity_score": 0.0403125521347703,
      "novelty_score": 0.984375,
      "recency_score": 0.8,
      "relevance_score": 0.39012302270183247,
      "bm25_score": 0.7267641902046712,
      "combined_score": 0.39012302270183247,
      "rank": 149
    },
    {
      "title": "MCNC: Manifold-Constrained Reparameterization for Neural Compression",
      "authors": [
        "Chayne Thrash",
        "Reed Andreas",
        "Ali Abbasi",
        "Parsa Nooralinejad",
        "Soroush Abbasi Koohpayegani",
        "Hamed Pirsiavash",
        "Soheil Kolouri"
      ],
      "abstract": "The outstanding performance of large foundational models across diverse tasks,\nfrom computer vision to speech and natural language processing, has significantly\nincreased their demand. However, storing and transmitting these models poses\nsignificant challenges due to their massive size (e.g., 750GB for Llama 3.1 405B).\nRecent literature has focused on compressing the original weights or reducing the\nnumber of parameters required for fine-tuning these models. These compression\nmethods generally constrain the parameter space, for example, through low-rank\nreparametrization (e.g., LoRA), pruning, or quantization (e.g., QLoRA) during\nor after the model training. In this paper, we present a novel model compres-\nsion method, which we term Manifold-Constrained Neural Compression (MCNC).\nThis method constrains the parameter space to low-dimensional pre-defined and\nfrozen nonlinear manifolds, which effectively cover this space. Given the preva-\nlence of good solutions in over-parameterized deep neural networks, we show that\nby constraining the parameter space to our proposed manifold, we can identify\nhigh-quality solutions while achieving unprecedented compression rates across\na wide variety of tasks and architectures. Through extensive experiments in\ncomputer vision and natural language processing tasks, we demonstrate that our\nmethod significantly outperforms state-of-the-art baselines in terms of compres-\nsion, accuracy, and/or model reconstruction time. Our code is publicly available at\nhttps://github.com/mint-vu/MCNC.",
      "url": "https://iclr.cc/virtual/2025/poster/29420",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=VMV8gefvq8",
      "citations": null,
      "categories": [],
      "id": "VMV8gefvq8",
      "track": "main",
      "status": "Poster",
      "keywords": "Model Compression;LoRA;PEFT;Transformers;ViT",
      "tldr": "",
      "primary_area": "other topics in machine learning (i.e., none of the above)",
      "similarity_score": 0.04727916054510844,
      "novelty_score": 0.9216867469879518,
      "recency_score": 0.9,
      "relevance_score": 0.39912121094156416,
      "bm25_score": 0.6831248759267722,
      "combined_score": 0.39912121094156416,
      "rank": 150
    },
    {
      "title": "Gradual Domain Adaptation via Manifold-Constrained Distributionally Robust Optimization",
      "authors": [
        "seyed amir hossein saberi",
        "Amir Najafi",
        "Amin Behjati",
        "Ala Emrani",
        "Yasaman Zolfimoselo",
        "Mahdi Shadrooy",
        "Abolfazl Motahari",
        "Babak Khalaj"
      ],
      "abstract": "The aim of this paper is to address the challenge of gradual domain adaptation within a class of manifold-constrained data distributions. In particular, we consider a sequence of $T\\ge2$ data distributions $P_1,\\ldots,P_T$ undergoing a gradual shift, where each pair of consecutive measures $P_i,P_{i+1}$ are close to each other in Wasserstein distance. We have a supervised dataset of size $n$ sampled from $P_0$, while for the subsequent distributions in the sequence, only unlabeled i.i.d. samples are available. Moreover, we assume that all distributions exhibit a known favorable attribute, such as (but not limited to) having intra-class soft/hard margins. In this context, we propose a methodology rooted in Distributionally Robust Optimization (DRO) with an adaptive Wasserstein radius. We theoretically show that this method guarantees the classification error across all $P_i$s can be suitably bounded. Our bounds rely on a newly introduced {\\it {compatibility}} measure, which fully characterizes the error propagation dynamics along the sequence. Specifically, for inadequately constrained distributions, the error can exponentially escalate as we progress through the gradual shifts. Conversely, for appropriately constrained distributions, the error can be demonstrated to be linear or even entirely eradicated. We have substantiated our theoretical findings through several experimental results.",
      "url": "https://neurips.cc/virtual/2024/poster/94967",
      "year": 2024,
      "venue": "NIPS 2024",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=UTNZKl5BUc",
      "citations": null,
      "categories": [],
      "id": "UTNZKl5BUc",
      "track": "main",
      "status": "Poster",
      "keywords": "Gradual Domain Adaptation;Distributionally Robust Optimization;Generalization Bound;Error Propagation Characterization",
      "tldr": "",
      "primary_area": "learning_theory",
      "similarity_score": 0.05476888698210291,
      "novelty_score": 0.9776357827476039,
      "recency_score": 0.8,
      "relevance_score": 0.3795080283741583,
      "bm25_score": 0.6769245409317581,
      "combined_score": 0.3795080283741583,
      "rank": 151
    },
    {
      "title": "Derivative-Free Diffusion Manifold-Constrained Gradient for Unified XAI",
      "authors": [
        "Won Jun Kim",
        "Hyungjin Chung",
        "Jaemin Kim",
        "Sangmin Lee",
        "Byeongsu Sim",
        "Jong Chul Ye"
      ],
      "abstract": "Gradient-based methods are a prototypical family of \"explainability for AI\" (XAI) techniques, especially for image-based models. However, they (1) require white-box access to models, (2) are vulnerable to adversarial attacks, and (3) produce attributions that lie off the image manifold, leading to explanations that are not amenable to human perception. To overcome these challenges, we introduce Derivative-Free Diffusion Manifold-Contrained Gradients (FreeMCG): by leveraging ensemble Kalman filters and diffusion models, we derive a derivative-free approximation of the model's gradient projected onto the data manifold, requiring access only to the model's outputs (i.e., black-box setting). We demonstrate the effectiveness of FreeMCG by applying it to both counterfactual generation and feature attribution, which have traditionally been treated as different tasks requiring distinct methods. Through comprehensive evaluation on both counterfactual explanation and feature attribution we show that our method yields state-of-the-art results for both tasks while preserving the essential properties expected of XAI tools. Code: https://github.com/one-june/FreeMCG.",
      "url": "https://cvpr.thecvf.com/virtual/2025/poster/32385",
      "year": 2025,
      "venue": "CVPR 2025",
      "source": "offline_cvpr",
      "doi": null,
      "pdf_url": "https://openaccess.thecvf.com/content/CVPR2025/papers/Kim_Derivative-Free_Diffusion_Manifold-Constrained_Gradient_for_Unified_XAI_CVPR_2025_paper.pdf",
      "citations": null,
      "categories": [],
      "id": "32385",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 0.052095246751324265,
      "novelty_score": 0.9248704663212436,
      "recency_score": 0.9,
      "relevance_score": 0.3954306439811383,
      "bm25_score": 0.6660068998524701,
      "combined_score": 0.3954306439811383,
      "rank": 152
    },
    {
      "title": "Manifold-Constrained Gaussian Process Inference for One-shot Learning of Unknown Ordinary Differential Equations",
      "authors": [],
      "abstract": "Learning unknown ordinary differential equations (ODEs) from a single trajectory of scarce, noisy data is challenging, especially with partial observability. We introduce MAGI-X, an integration-free framework that couples a neural vector field with a Gaussian process prior over trajectories and enforces ODE consistency via a GP manifold constraint, thereby circumventing traditional numerical integration. Across canonical examples (FitzHugh--Nagumo, Lotka--Volterra, and Hes1), MAGI-X achieves better accuracy in both fitting and forecasting while requiring comparable or less computation time than benchmark methods NPODE and Neural ODE, with runtime scaling linearly in state dimension. MAGI-X offers a practical solution for \\emph{partially observed} systems without bespoke priors or imputation heuristics, where existing methods struggle. The GP posterior further yields calibrated uncertainty, and experiments demonstrate robustness across initial conditions. We show practicality on seasonal flu data with rolling multi-week forecasts from noisy signals.\n These properties establish MAGI-X as a fast, accurate, and robust tool for data-driven discovery of nonlinear dynamics from a single noisy trajectory.",
      "url": "https://openreview.net/forum?id=ECc2td0LCZ",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "ECc2td0LCZ",
      "track": "main",
      "status": "Active",
      "keywords": "Gaussian Process;ODE learning",
      "tldr": "",
      "primary_area": "applications to physical sciences (physics, chemistry, biology, etc.)",
      "similarity_score": 0.029942329949874558,
      "novelty_score": 0.9296202531645569,
      "recency_score": 1.0,
      "relevance_score": 0.4016478470691398,
      "bm25_score": 0.6422171602805914,
      "combined_score": 0.4016478470691398,
      "rank": 153
    },
    {
      "title": "CFG++: Manifold-constrained Classifier Free Guidance for Diffusion Models",
      "authors": [
        "Hyungjin Chung",
        "Jeongsol Kim",
        "Geon Yeong Park",
        "Hyelin Nam",
        "Jong Chul Ye"
      ],
      "abstract": "Classifier-free guidance (CFG) is a fundamental tool in modern diffusion models for text-guided generation. Although effective, CFG has notable drawbacks. For instance, DDIM with CFG lacks invertibility, complicating image editing; furthermore, high guidance scales, essential for high-quality outputs, frequently result in issues like mode collapse. Contrary to the widespread belief that these are inherent limitations of diffusion models, this paper reveals that the problems actually stem from the off-manifold phenomenon associated with CFG, rather than the diffusion models themselves. More specifically, inspired by the recent advancements of diffusion model-based inverse problem solvers (DIS),  we reformulate text-guidance as an inverse problem with a text-conditioned score matching loss and develop CFG++, a novel approach that tackles the off-manifold challenges inherent in traditional CFG. CFG++ features a surprisingly simple fix to CFG, yet it offers significant improvements, including better sample quality for text-to-image generation, invertibility, smaller guidance scales,  reduced etc. Furthermore, CFG++ enables seamless interpolation between unconditional and conditional sampling at lower guidance scales, consistently outperforming traditional CFG at all scales. Moreover, CFG++ can be easily integrated into the high-order diffusion solvers and naturally extends to distilled diffusion models. Experimental results confirm that our method significantly enhances performance in text-to-image generation, DDIM inversion, editing, and solving inverse problems, suggesting a wide-ranging impact and potential applications in various fields that utilize text guidance. Project Page: https://cfgpp-diffusion.github.io/anon",
      "url": "https://iclr.cc/virtual/2025/poster/30421",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=E77uvbOTtp",
      "citations": null,
      "categories": [],
      "id": "E77uvbOTtp",
      "track": "main",
      "status": "Poster",
      "keywords": "Diffusion models;Manifold;Classifier-free guidance",
      "tldr": "",
      "primary_area": "generative models",
      "similarity_score": 0.022999950287026,
      "novelty_score": 0.935064935064935,
      "recency_score": 0.9,
      "relevance_score": 0.367795805946054,
      "bm25_score": 0.6029860695331539,
      "combined_score": 0.367795805946054,
      "rank": 154
    },
    {
      "title": "Escaping from saddle points on Riemannian manifolds",
      "authors": [
        "Yue Sun",
        "Nicolas Flammarion",
        "Maryam Fazel"
      ],
      "abstract": "We consider minimizing a nonconvex, smooth function $f$ on a Riemannian manifold $\\mathcal{M}$. We show that a perturbed version of the gradient descent algorithm converges to a second-order stationary point for this problem (and hence is able to escape saddle points on the manifold). While the unconstrained problem is well-studied, our result is the first to prove such a rate for  nonconvex, manifold-constrained problems.\nThe rate of convergence depends as $1/\\epsilon^2$ on the accuracy $\\epsilon$, which matches a rate known only for unconstrained smooth minimization. The convergence rate also has a  polynomial dependence on the parameters denoting the curvature of the manifold and the smoothness of the function.",
      "url": "https://nips.cc/virtual/2019/poster/13817",
      "year": 2019,
      "venue": "NIPS 2019",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2019/file/24e01830d213d75deb99c22b9cd91ddd-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "13817",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 0.0727753310932536,
      "novelty_score": 0.9562499999999999,
      "recency_score": 0.30000000000000004,
      "relevance_score": 0.2192030040168738,
      "bm25_score": 0.4579013489629923,
      "combined_score": 0.2192030040168738,
      "rank": 155
    },
    {
      "title": "Trajectory Optimization on Manifolds: A Theoretically-Guaranteed Embedded Sequential Convex Programming Approach",
      "authors": [
        "Riccardo Bonalli",
        "Abhishek Cauligi",
        "Andrew Bylard",
        "Thomas Lew",
        "Marco Pavone"
      ],
      "abstract": "Sequential Convex Programming (SCP) has recently gained popularity as a tool for trajectory optimization due to its sound theoretical properties and practical performance. Yet, most SCP-based methods for trajectory optimization are restricted to Euclidean settings, which precludes their application to problem instances where one must reason about manifold-type constraints (that is, constraints, such as loop closure, which restrict the motion of a system to a subset of the ambient space). The aim of this paper is to fill this gap by extending SCP-based trajectory optimization methods to a manifold setting. The key insight is to leverage geometric embeddings to lift a manifold-constrained trajectory optimization problem into an equivalent problem defined over a space enjoying a Euclidean structure. This insight allows one to extend existing SCP methods to a manifold setting in a fairly natural way. In particular, we present a SCP algorithm for manifold problems with refined theoretical guarantees that resemble those derived for the Euclidean setting, and demonstrate its practical performance via numerical experiments.",
      "url": "https://www.roboticsproceedings.org/rss15/p78.html",
      "year": 2019,
      "venue": "RSS 2019",
      "source": "offline_rss",
      "doi": null,
      "pdf_url": "https://www.roboticsproceedings.org/rss15/p78.pdf",
      "citations": null,
      "categories": [],
      "id": "569fe8abe2",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 0.05893370730515055,
      "novelty_score": 0.9551645856980704,
      "recency_score": 0.30000000000000004,
      "relevance_score": 0.1932841352913623,
      "bm25_score": 0.38534674366605703,
      "combined_score": 0.1932841352913623,
      "rank": 156
    },
    {
      "title": "DGSolver: Diffusion Generalist Solver with Universal Posterior Sampling for Image Restoration",
      "authors": [
        "Hebaixu Wang",
        "Jing Zhang",
        "Haonan Guo",
        "Di Wang",
        "Jiayi Ma",
        "Bo Du"
      ],
      "abstract": "Diffusion models have achieved remarkable progress in universal image restoration. However, existing methods perform naive inference in the reverse process, which leads to cumulative errors under limited sampling steps and large step intervals. Moreover, they struggle to balance the commonality of degradation representations with restoration quality, often depending on complex compensation mechanisms that enhance fidelity at the expense of efficiency. To address these challenges, we introduce \\textbf{DGSolver}, a diffusion generalist solver with universal posterior sampling. We first derive the exact ordinary differential equations for generalist diffusion models to unify degradation representations and design tailored high-order solvers with a queue-based accelerated sampling strategy to improve both accuracy and efficiency. We then integrate universal posterior sampling to better approximate manifold-constrained gradients, yielding a more accurate noise estimation and correcting errors in inverse inference. Extensive experiments demonstrate that DGSolver outperforms state-of-the-art methods in restoration accuracy, stability, and scalability, both qualitatively and quantitatively. Code and models are publicly available at https://github.com/MiliLab/DGSolver.",
      "url": "https://openreview.net/forum?id=ghhKZ0NaQN",
      "year": 2025,
      "venue": "NIPS 2025",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "ghhKZ0NaQN",
      "track": "main",
      "status": "Poster",
      "keywords": "Image restoration;diffusion generalist solver;universal posterior sampling;deep learning",
      "tldr": "",
      "primary_area": "deep_learning",
      "similarity_score": 0.017664586819726297,
      "novelty_score": 0.9412751677852348,
      "recency_score": 0.9,
      "relevance_score": 0.2993322218908988,
      "bm25_score": 0.3801094861499363,
      "combined_score": 0.2993322218908988,
      "rank": 157
    },
    {
      "title": "Riemannian batch normalization for SPD neural networks",
      "authors": [
        "Daniel Brooks",
        "Olivier Schwander",
        "Frederic Barbaresco",
        "Jean-Yves Schneider",
        "Matthieu Cord"
      ],
      "abstract": "Covariance matrices have attracted attention for machine learning applications due\nto their capacity to capture interesting structure in the data. The main challenge\nis that one needs to take into account the particular geometry of the Riemannian\nmanifold of symmetric positive definite (SPD) matrices they belong to. In the con-\ntext of deep networks, several architectures for these matrices have recently been\nproposed. In our article, we introduce a Riemannian batch normalization (batch-\nnorm) algorithm, which generalizes the one used in Euclidean nets. This novel\nlayer makes use of geometric operations on the manifold, notably the Riemannian\nbarycenter, parallel transport and non-linear structured matrix transformations. We\nderive a new manifold-constrained gradient descent algorithm working in the space\nof SPD matrices, allowing to learn the batchnorm layer. We validate our proposed\napproach with experiments in three different contexts on diverse data types: a\ndrone recognition dataset from radar observations, and on emotion and action\nrecognition datasets from video and motion capture data. Experiments show that\nthe Riemannian batchnorm systematically gives better classification performance\ncompared with leading methods and a remarkable robustness to lack of data.",
      "url": "https://nips.cc/virtual/2019/poster/14444",
      "year": 2019,
      "venue": "NIPS 2019",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://papers.nips.cc/paper_files/paper/2019/file/6e69ebbfad976d4637bb4b39de261bf7-Paper.pdf",
      "citations": null,
      "categories": [],
      "id": "14444",
      "track": "main",
      "status": "Poster",
      "keywords": "",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 0.0406106554688012,
      "novelty_score": 0.9296875,
      "recency_score": 0.30000000000000004,
      "relevance_score": 0.18595832237501342,
      "bm25_score": 0.37925041911457685,
      "combined_score": 0.18595832237501342,
      "rank": 158
    },
    {
      "title": "Federated Dynamical Low-Rank Training with Global Loss Convergence Guarantees",
      "authors": [
        "Steffen Schotthöfer",
        "M. Paul Laiu"
      ],
      "abstract": "We propose a federated dynamical low-rank training (FeDLRT) scheme to reduce client compute and communication costs - two significant performance bottlenecks in horizontal federated learning. Our method builds upon dynamical low-rank splitting schemes for manifold-constrained optimization to create a global low-rank basis of network weights, which enables client training on a small coefficient matrix. This global low-rank basis that allows us to incorporate a variance correction scheme and prove global loss descent and convergence to a stationary point. FeDLRT features  dynamic augmentation and truncation of the low-rank bases to optimize computing and communication resource utilization. Notably FeDLRT only trains a small coefficient matrix per client. We demonstrate the efficiency of FeDLRT in an array of computer vision benchmarks with both i.i.d. and non-i.i.d. data distributions and show a reduction of client compute and communication costs by up to an order of magnitude with minimal impacts on global accuracy.\nFeDLRT performs as well as classical methods such as FedAvg and FedLin, with a fraction of the memory and compute requirements.",
      "url": "https://openreview.net/forum?id=RAC3ng3TSN",
      "year": 2025,
      "venue": "ICLR 2025",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "RAC3ng3TSN",
      "track": "main",
      "status": "Reject",
      "keywords": "Federated Learning;Low-Rank;Model Compression;Efficient Federated Learning",
      "tldr": "",
      "primary_area": "optimization",
      "similarity_score": 0.016173397434332607,
      "novelty_score": 0.972457627118644,
      "recency_score": 0.9,
      "relevance_score": 0.2983705871486929,
      "bm25_score": 0.3783952263946436,
      "combined_score": 0.2983705871486929,
      "rank": 159
    },
    {
      "title": "Beyond Minimax: Structure-Aware Learning for Differential Games",
      "authors": [],
      "abstract": "A central challenge in artificial intelligence is to design agents that solve structured engineering problems, such as zero-sum differential games, without handcrafted solutions or expert demonstrations. Differential games capture multi-agent interactions with opposing objectives, where optimal strategies are defined by equilibrium conditions. Classical theory based on Pontryagin’s Maximum Principle (PMP) and the Hamilton--Jacobi--Isaacs (HJI) equations provides principled foundations, but these conditions are rarely tractable in practice. Deep learning, by contrast, offers flexible function approximation but typically ignores such structure and depends on large datasets or extensive online interactions.\n\nWe introduce a framework that embeds equilibrium conditions and terminal constraints from the calculus of variations directly into the training objective. This enables neural networks to jointly learn state, control, and costate trajectories while handling variable terminal times and manifold-constrained terminal states, yielding approximate saddle-point equilibria. We illustrate our approach with the pursuit--evasion game \\emph{Lady in the Lake}, showing that our method recovers structural properties of analytical solutions and generalizes to novel scenarios without supervision, pointing toward principled, structure-aware deep models for solving previously intractable differential games.",
      "url": "https://openreview.net/forum?id=61jN0L0aoJ",
      "year": 2026,
      "venue": "ICLR 2026",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "61jN0L0aoJ",
      "track": "main",
      "status": "Active",
      "keywords": "pursuit-evasion game;calculus of variations;pontryagin's maximum principle",
      "tldr": "",
      "primary_area": "neurosymbolic & hybrid AI systems (physics-informed, logic & formal reasoning, etc.)",
      "similarity_score": 0.01895355849423907,
      "novelty_score": 0.9329268292682927,
      "recency_score": 1.0,
      "relevance_score": 0.3186949754087247,
      "bm25_score": 0.3766963595348432,
      "combined_score": 0.3186949754087247,
      "rank": 160
    },
    {
      "title": "Diffusion Probabilistic Models for Structured Node Classification",
      "authors": [
        "Hyosoon Jang",
        "Seonghyun Park",
        "Sangwoo Mo",
        "Sungsoo Ahn"
      ],
      "abstract": "This paper studies structured node classification on graphs, where the predictions should consider dependencies between the node labels. In particular, we focus on solving the problem for partially labeled graphs where it is essential to incorporate the information in the known label for predicting the unknown labels. To address this issue, we propose a novel framework leveraging the diffusion probabilistic model for structured node classification (DPM-SNC). At the heart of our framework is the extraordinary capability of DPM-SNC to (a) learn a joint distribution over the labels with an expressive reverse diffusion process and (b) make predictions conditioned on the known labels utilizing manifold-constrained sampling. Since the DPMs lack training algorithms for partially labeled data, we design a novel training algorithm to apply DPMs, maximizing a new variational lower bound. We also theoretically analyze how DPMs benefit node classification by enhancing the expressive power of GNNs based on proposing AGG-WL, which is strictly more powerful than the classic 1-WL test. We extensively verify the superiority of our DPM-SNC in diverse scenarios, which include not only the transductive setting on partially labeled graphs but also the inductive setting and unlabeled graphs.",
      "url": "https://nips.cc/virtual/2023/poster/72405",
      "year": 2023,
      "venue": "NIPS 2023",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=CxUuCydMDU",
      "citations": null,
      "categories": [],
      "id": "CxUuCydMDU",
      "track": "main",
      "status": "Poster",
      "keywords": "diffusion model;graph neural network;structured prediction;node classification",
      "tldr": "",
      "primary_area": "",
      "similarity_score": 0.016848273089726352,
      "novelty_score": 0.9296875,
      "recency_score": 0.7,
      "relevance_score": 0.2543820603168867,
      "bm25_score": 0.36442526129989616,
      "combined_score": 0.2543820603168867,
      "rank": 161
    },
    {
      "title": "Toward a Unified Geometry Understanding : Riemannian Diffusion Framework for Graph Generation and Prediction",
      "authors": [
        "Yisen Gao",
        "Xingcheng Fu",
        "Qingyun Sun",
        "Jianxin Li",
        "Xianxian LI"
      ],
      "abstract": "Graph diffusion models have made significant progress in learning structured graph data and have demonstrated strong potential for predictive tasks. Existing approaches typically embed node, edge, and graph-level features into a unified latent space, modeling prediction tasks including classification and regression as a form of conditional generation.  However, due to the non-Euclidean nature of graph data, features of different curvatures are entangled in the same latent space without releasing their geometric potential. To address this issue, we aim to construt an ideal Riemannian diffusion model to capture distinct manifold signatures of complex graph data and learn their distribution.  This goal faces two challenges: numerical instability caused by exponential mapping during the encoding proces and manifold deviation during diffusion generation.  To address these challenges, we propose **GeoMancer**: a novel Riemannian graph diffusion framework for both generation and prediction tasks. To mitigate numerical instability, we replace exponential mapping with an isometric-invariant Riemannian gyrokernel approach and decouple multi-level features onto their respective task-specific manifolds to learn optimal representations.  To address manifold deviation, we introduce a manifold-constrained diffusion method and a self-guided strategy for unconditional generation, ensuring that the generated data remains aligned with the manifold signature. Extensive experiments validate the effectiveness of our approach, demonstrating superior performance across a variety of tasks.",
      "url": "https://openreview.net/forum?id=QQqDBRRslp",
      "year": 2025,
      "venue": "NIPS 2025",
      "source": "offline_nips",
      "doi": null,
      "pdf_url": "",
      "citations": null,
      "categories": [],
      "id": "QQqDBRRslp",
      "track": "main",
      "status": "Poster",
      "keywords": "Graph Generation;Hyperbolic Graph Learning;Riemannian Manifold;Graph Learning",
      "tldr": "",
      "primary_area": "deep_learning",
      "similarity_score": 0.051566305159198356,
      "novelty_score": 0.9201030927835052,
      "recency_score": 0.9,
      "relevance_score": 0.296059438662671,
      "bm25_score": 0.3352984903830381,
      "combined_score": 0.296059438662671,
      "rank": 162
    },
    {
      "title": "Decomposed Diffusion Sampler for Accelerating Large-Scale Inverse Problems",
      "authors": [
        "Hyungjin Chung",
        "Suhyeon Lee",
        "Jong Chul Ye"
      ],
      "abstract": "Krylov subspace, which is generated by multiplying a given vector by the matrix of a linear transformation and its successive powers,  has been extensively studied in classical optimization literature to design algorithms that converge quickly for large linear inverse problems. For example, the conjugate gradient method (CG), one of the most popular Krylov subspace methods, is based on the idea of minimizing the residual error in the Krylov subspace. However, with the recent advancement of high-performance diffusion solvers for inverse problems, it is not clear how classical wisdom can be synergistically combined with modern diffusion models. In this study, we propose a novel and efficient diffusion sampling strategy that synergistically combines the diffusion sampling and Krylov subspace methods. Specifically, we prove that if the tangent space at a denoised sample by Tweedie's formula forms a  Krylov subspace, then the CG initialized with the denoised data ensures the data consistency update to remain in the tangent space. This negates the need to compute the manifold-constrained gradient (MCG), leading to a more efficient diffusion sampling method. Our method is applicable regardless of the parametrization and setting (i.e., VE, VP). Notably, we achieve state-of-the-art reconstruction quality on challenging real-world medical inverse imaging problems, including multi-coil MRI reconstruction and 3D CT reconstruction. Moreover, our proposed method achieves more than 80 times faster inference time than the previous state-of-the-art method. Code is available at https://github.com/HJ-harry/DDS",
      "url": "https://iclr.cc/virtual/2024/poster/19121",
      "year": 2024,
      "venue": "ICLR 2024",
      "source": "offline_iclr",
      "doi": null,
      "pdf_url": "https://openreview.net/pdf?id=DsEhqQtfAG",
      "citations": null,
      "categories": [],
      "id": "DsEhqQtfAG",
      "track": "main",
      "status": "Poster",
      "keywords": "Diffusion models; Inverse problems; Krylov subspace",
      "tldr": "",
      "primary_area": "generative models",
      "similarity_score": 0.016094456133094376,
      "novelty_score": 0.9393706830391404,
      "recency_score": 0.8,
      "relevance_score": 0.2648185037127727,
      "bm25_score": 0.3333005562428146,
      "combined_score": 0.2648185037127727,
      "rank": 163
    },
    {
      "title": "Hyper-Graph Regularized Constrained NMF for Selecting Differentially Expressed Genes and Tumor Classification",
      "authors": [
        "Cui-Na Jiao",
        "Ying-Lian Gao",
        "Na Yu",
        "Jin-Xing Liu",
        "Lianyong Qi"
      ],
      "abstract": "Non-negative Matrix Factorization (NMF) is a dimensionality reduction approach for learning a parts-based and linear representation of non-negative data. It has attracted more attention because of that. In practice, NMF not only neglects the manifold structure of data samples, but also overlooks the priori label information of different classes. In this paper, a novel matrix decomposition method called Hyper-graph regularized Constrained Non-negative Matrix Factorization (HCNMF) is proposed for selecting differentially expressed genes and tumor sample classification. The advantage of hyper-graph learning is to capture local spatial information in high dimensional data. This method incorporates a hyper-graph regularization constraint to consider the higher order data sample relationships. The application of hyper-graph theory can effectively find pathogenic genes in cancer datasets. Besides, the label information is further incorporated in the objective function to improve the discriminative ability of the decomposition matrix. Supervised learning with label information greatly improves the classification effect. We also provide the iterative update rules and convergence proofs for the optimization problems of HCNMF. Experiments under The Cancer Genome Atlas (TCGA) datasets confirm the superiority of HCNMF algorithm compared with other representative algorithms through a set of evaluations.",
      "url": "https://www.semanticscholar.org/paper/f68b6e6c27a63f965cd6612cfd67acf1b40d3b5b",
      "year": 2020,
      "venue": "IEEE journal of biomedical and health informatics",
      "source": "semantic_scholar",
      "doi": "10.1109/JBHI.2020.2975199",
      "pdf_url": "",
      "citations": 46,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.1310448288283798,
      "novelty_score": 0.934351145038168,
      "recency_score": 0.4,
      "relevance_score": 0.13771344864851393,
      "bm25_score": 0.0,
      "combined_score": 0.13771344864851393,
      "rank": 164
    },
    {
      "title": "Ground states for the planar NLSE with a point defect as minimizers of the constrained energy",
      "authors": [
        "R. Adami",
        "F. Boni",
        "R. Carlone",
        "L. Tentarelli"
      ],
      "abstract": "We investigate the ground states for the focusing, subcritical nonlinear Schrödinger equation with a point defect in dimension two, defined as the minimizers of the energy functional at fixed mass. We prove that ground states exist for every positive mass and show a logarithmic singularity at the defect. Moreover, up to a multiplication by a constant phase, they are positive, radially symmetric, and decreasing along the radial direction. In order to overcome the obstacles arising from the uncommon structure of the energy space, that complicates the application of standard rearrangement theory, we move to the study of the minimizers of the action functional on the Nehari manifold and then establish a connection with the original problem. A refinement of a classical result on rearrangements is proved to obtain qualitative features of the ground states.",
      "url": "https://www.semanticscholar.org/paper/a25875f86614ddf77564ca2f0be903da9dd8d583",
      "year": 2021,
      "venue": "Calculus of Variations and Partial Differential Equations",
      "source": "semantic_scholar",
      "doi": "10.1007/s00526-022-02310-8",
      "pdf_url": "",
      "citations": 32,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.023660427396043703,
      "novelty_score": 0.9073387694588584,
      "recency_score": 0.5,
      "relevance_score": 0.11989812821881313,
      "bm25_score": 0.0,
      "combined_score": 0.11989812821881313,
      "rank": 165
    },
    {
      "title": "The non-adiabatic nanoreactor: towards the automated discovery of photochemistry†",
      "authors": [
        "Elisa Pieri",
        "Dean Lahana",
        "Alexander M Chang",
        "Cody R Aldaz",
        "K. Thompson",
        "T. Martínez"
      ],
      "abstract": "The ab initio nanoreactor has previously been introduced to automate reaction discovery for ground state chemistry. In this work, we present the nonadiabatic nanoreactor, an analogous framework for excited state reaction discovery. We automate the study of nonadiabatic decay mechanisms of molecules by probing the intersection seam between adiabatic electronic states with hyper-real metadynamics, sampling the branching plane for relevant conical intersections, and performing seam-constrained path searches. We illustrate the effectiveness of the nonadiabatic nanoreactor by applying it to benzene, a molecule with rich photochemistry and a wide array of photochemical products. Our study confirms the existence of several types of S0/S1 and S1/S2 conical intersections which mediate access to a variety of ground state stationary points. We elucidate the connections between conical intersection energy/topography and the resulting photoproduct distribution, which changes smoothly along seam space segments. The exploration is performed with minimal user input, and the protocol requires no previous knowledge of the photochemical behavior of a target molecule. We demonstrate that the nonadiabatic nanoreactor is a valuable tool for the automated exploration of photochemical reactions and their mechanisms.",
      "url": "https://www.semanticscholar.org/paper/5968f7fd9a1e57e9fc713d17a02013297599a842",
      "year": 2021,
      "venue": "Chemical Science",
      "source": "semantic_scholar",
      "doi": "10.1039/d1sc00775k",
      "pdf_url": "https://pubs.rsc.org/en/content/articlepdf/2021/sc/d1sc00775k",
      "citations": 24,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.05555078560262291,
      "novelty_score": 0.9173884077281811,
      "recency_score": 0.5,
      "relevance_score": 0.12626523568078687,
      "bm25_score": 0.0,
      "combined_score": 0.12626523568078687,
      "rank": 166
    },
    {
      "title": "MoDANet: Multi-Task Deep Network for Joint Automatic Modulation Classification and Direction of Arrival Estimation",
      "authors": [
        "Van-Sang Doan",
        "Thien Huynh-The",
        "Van-Phuc Hoang",
        "Duy T. Nguyen"
      ],
      "abstract": "In this letter, a multi-task deep convolutional neural network, namely MoDANet, is proposed to perform modulation classification and DOA estimation simultaneously. In particular, the network architecture is designed with multiple residual modules, which tackle the vanishing gradient problem. The multi-task learning (MTL) efficiency of MoDANet was evaluated with different variants of Y-shaped connection and fine-tuning some hyper-parameters of the deep network. As a result, MoDANet with one shared residual module using more filters, larger filter size, and longer signal length can achieve better performance of modulation classification and DOA estimation, but those might result in higher computational complexity. Therefore, choosing these parameters to attain a good trade-off between accuracy and computational cost is important, especially for resource-constrained devices. The network is investigated with two typical propagation channel models, including Pedestrian A and Vehicular A, to show the effect of those channels on the efficiency of the network. Remarkably, our work is the first DL-based MTL model to handle two unrelated tasks of modulation classification and DOA estimation.",
      "url": "https://www.semanticscholar.org/paper/816c248b74fbd8db43a85747dde57beedab2dbaf",
      "year": 2022,
      "venue": "IEEE Communications Letters",
      "source": "semantic_scholar",
      "doi": "10.1109/lcomm.2021.3132018",
      "pdf_url": "",
      "citations": 22,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.03193742649046898,
      "novelty_score": 0.9228780697975011,
      "recency_score": 0.6,
      "relevance_score": 0.1383812279471407,
      "bm25_score": 0.0,
      "combined_score": 0.1383812279471407,
      "rank": 167
    },
    {
      "title": "SLM: A Smoothed First-Order Lagrangian Method for Structured Constrained Nonconvex Optimization",
      "authors": [
        "Songtao Lu"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/245e67d4acd761b1d5b82bd5279803a48831c1bf",
      "year": 2023,
      "venue": "Neural Information Processing Systems",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 19,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.047901740807890614,
      "novelty_score": 0.9334442595673877,
      "recency_score": 0.7,
      "relevance_score": 0.16197052224236716,
      "bm25_score": 0.0,
      "combined_score": 0.16197052224236716,
      "rank": 168
    },
    {
      "title": "ProxNLP: a primal-dual augmented Lagrangian solver for nonlinear programming in Robotics and beyond",
      "authors": [
        "Wilson Jallet",
        "Antoine Bambade",
        "N. Mansard",
        "Justin Carpentier"
      ],
      "abstract": "Mathematical optimization is the workhorse behind several aspects of modern robotics and control. In these applications, the focus is on constrained optimization, and the ability to work on manifolds (such as the classical matrix Lie groups), along with a specific requirement for robustness and speed. In recent years, augmented Lagrangian methods have seen a resurgence due to their robustness and flexibility, their connections to (inexact) proximal-point methods, and their interoperability with Newton or semismooth Newton methods. In the sequel, we present primal-dual augmented Lagrangian method for inequality-constrained problems on manifolds, which we introduced in our recent work, as well as an efficient C++ implementation suitable for use in robotics applications and beyond.",
      "url": "https://www.semanticscholar.org/paper/9a59529464edf5dc477afd2dcce76b892b2964f1",
      "year": 2022,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2210.02109",
      "pdf_url": "http://arxiv.org/pdf/2210.02109",
      "citations": 18,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.07073186916795456,
      "novelty_score": 0.9157458563535913,
      "recency_score": 0.6,
      "relevance_score": 0.14841956075038637,
      "bm25_score": 0.0,
      "combined_score": 0.14841956075038637,
      "rank": 169
    },
    {
      "title": "Asymptotic profiles for a nonlinear Schrödinger equation with critical combined powers nonlinearity",
      "authors": [
        "Shiwang Ma",
        "Vitaly Moroz"
      ],
      "abstract": "We study asymptotic behaviour of positive ground state solutions of the nonlinear Schrödinger equation $$\\begin{aligned} -\\Delta u+u=u^{2^*-1}+\\lambda u^{q-1} \\quad \\textrm{in}\\, {\\mathbb {R}}^N,\\qquad \\qquad \\qquad \\qquad \\qquad {(P_\\lambda )} \\end{aligned}$$ - Δ u + u = u 2 ∗ - 1 + λ u q - 1 in R N , ( P λ ) where $$N\\ge 3$$ N ≥ 3 is an integer, $$2^{*}=\\frac{2N}{N-2}$$ 2 ∗ = 2 N N - 2 is the Sobolev critical exponent, $$20$$ λ > 0 is a parameter. It is known that as $$\\lambda \\rightarrow 0$$ λ → 0 , after a rescaling the ground state solutions of $$(P_\\lambda )$$ ( P λ ) converge to a particular solution of the critical Emden-Fowler equation $$-\\Delta u=u^{2^*-1}$$ - Δ u = u 2 ∗ - 1 . We establish a novel sharp asymptotic characterisation of such a rescaling, which depends in a non-trivial way on the space dimension $$N=3$$ N = 3 , $$N=4$$ N = 4 or $$N \\ge 5$$ N ≥ 5 . We also discuss a connection of these results with a mass constrained problem associated to $$(P_{\\lambda })$$ ( P λ ) . Unlike previous work of this type, our method is based on the Nehari-Pohožaev manifold minimization, which allows to control the $$L^{2}$$ L 2 norm of the groundstates.",
      "url": "https://www.semanticscholar.org/paper/9b75f0de537b0eb389208a43392d8fe2424a38cf",
      "year": 2023,
      "venue": "Mathematische Zeitschrift",
      "source": "semantic_scholar",
      "doi": "10.1007/s00209-023-03271-0",
      "pdf_url": "https://link.springer.com/content/pdf/10.1007/s00209-023-03271-0.pdf",
      "citations": 11,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.018014931534182297,
      "novelty_score": 0.9367346938775512,
      "recency_score": 0.7,
      "relevance_score": 0.14980447946025466,
      "bm25_score": 0.0,
      "combined_score": 0.14980447946025466,
      "rank": 170
    },
    {
      "title": "Nonholonomic and constrained variational mechanics",
      "authors": [
        "A. D. Lewis"
      ],
      "abstract": "Equations governing mechanical systems with nonholonomic constraints can be developed in two ways: (1) using the physical principles of Newtonian mechanics; (2) using a constrained variational principle. Generally, the two sets of resulting equations are not equivalent. While mechanics arises from the first of these methods, sub-Riemannian geometry is a special case of the second. Thus both sets of equations are of independent interest. The equations in both cases are carefully derived using a novel Sobolev analysis where infinite-dimensional Hilbert manifolds are replaced with infinite-dimensional Hilbert spaces for the purposes of analysis. A useful representation of these equations is given using the so-called constrained connection derived from the system's Riemannian metric, and the constraint distribution and its orthogonal complement. In the special case of sub-Riemannian geometry, some observations are made about the affine connection formulation of the equations for extremals. Using the affine connection formulation of the equations, the physical and variational equations are compared and conditions are given that characterise when all physical solutions arise as extremals in the variational formulation. The characterisation is complete in the real analytic case, while in the smooth case a locally constant rank assumption must be made. The main construction is that of the largest affine subbundle variety of a subbundle that is invariant under the flow of an affine vector field on the total space of a vector bundle.",
      "url": "https://www.semanticscholar.org/paper/69a2d7a0fccd5b709fe2aeedfad1e303adec446d",
      "year": 2020,
      "venue": "",
      "source": "semantic_scholar",
      "doi": "10.3934/jgm.2020013",
      "pdf_url": "https://www.aimsciences.org/article/exportPdf?id=6438a283-c843-455a-9966-141eb920f0c5",
      "citations": 10,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.025217327934330836,
      "novelty_score": 0.9250302297460702,
      "recency_score": 0.4,
      "relevance_score": 0.09156519838029926,
      "bm25_score": 0.0,
      "combined_score": 0.09156519838029926,
      "rank": 171
    },
    {
      "title": "Cortico-Cerebellar Hyper-Connections and Reduced Purkinje Cells Behind Abnormal Eyeblink Conditioning in a Computational Model of Autism Spectrum Disorder",
      "authors": [
        "Emiliano Trimarco",
        "Pierandrea Mirino",
        "Daniele Caligiore"
      ],
      "abstract": "Empirical evidence suggests that children with autism spectrum disorder (ASD) show abnormal behavior during delay eyeblink conditioning. They show a higher conditioned response learning rate and earlier peak latency of the conditioned response signal. The neuronal mechanisms underlying this autistic behavioral phenotype are still unclear. Here, we use a physiologically constrained spiking neuron model of the cerebellar-cortical system to investigate which features are critical to explaining atypical learning in ASD. Significantly, the computer simulations run with the model suggest that the higher conditioned responses learning rate mainly depends on the reduced number of Purkinje cells. In contrast, the earlier peak latency mainly depends on the hyper-connections of the cerebellum with sensory and motor cortex. Notably, the model has been validated by reproducing the behavioral data collected from studies with real children. Overall, this article is a starting point to understanding the link between the behavioral and neurobiological basis in ASD learning. At the end of the paper, we discuss how this knowledge could be critical for devising new treatments.",
      "url": "https://www.semanticscholar.org/paper/105dfc9dd495e4e279cc08f6ae88cfa481806c1f",
      "year": 2021,
      "venue": "Frontiers in Systems Neuroscience",
      "source": "semantic_scholar",
      "doi": "10.3389/fnsys.2021.666649",
      "pdf_url": "https://doi.org/10.3389/fnsys.2021.666649",
      "citations": 7,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.11479335833679638,
      "novelty_score": 0.9392905866302864,
      "recency_score": 0.5,
      "relevance_score": 0.1372380075010389,
      "bm25_score": 0.0,
      "combined_score": 0.1372380075010389,
      "rank": 172
    },
    {
      "title": "Revealing the hidden structure of disordered materials by parameterizing their local structural manifold",
      "authors": [
        "Thomas J. Hardin",
        "Michael Chandross",
        "Rahul Meena",
        "Spencer Fajardo",
        "Dimitris G. Giovanis",
        "I. Kevrekidis",
        "Michael L. Falk",
        "Michael D. Shields"
      ],
      "abstract": "Durable interest in developing a framework for the detailed structure of glassy materials has produced numerous structural descriptors that trade off between general applicability and interpretability. However, none approach the combination of simplicity and wide-ranging predictive power of the lattice-grain-defect framework for crystalline materials. Working from the hypothesis that the local atomic environments of a glassy material are constrained by enthalpy minimization to a low-dimensional manifold in atomic coordinate space, we develop a generalized distance function, the Gaussian Integral Inner Product (GIIP) distance, in connection with agglomerative clustering and diffusion maps, to parameterize that manifold. Applying this approach to a two-dimensional model crystal and a three-dimensional binary model metallic glass results in parameters interpretable as coordination number, composition, volumetric strain, and local symmetry. In particular, we show that a more slowly quenched glass has a higher degree of local tetrahedral symmetry at the expense of cyclic symmetry. While these descriptors require post-hoc interpretation, they minimize bias rooted in crystalline materials science and illuminate a range of structural trends that might otherwise be missed. The structure of crystalline materials plays a central role in materials science, but the disordered structure of metallic glass is difficult to characterize and describe. Here, the authors use diffusion maps on atomistic data to obtain general structural descriptors tied to atomic positions.",
      "url": "https://www.semanticscholar.org/paper/7ddf9876b5e857d2fbbb010238501dd51550e605",
      "year": 2024,
      "venue": "Nature Communications",
      "source": "semantic_scholar",
      "doi": "10.1038/s41467-024-48449-0",
      "pdf_url": "https://www.nature.com/articles/s41467-024-48449-0.pdf",
      "citations": 6,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.031167078968806713,
      "novelty_score": 0.9471428571428572,
      "recency_score": 0.8,
      "relevance_score": 0.17175012369064205,
      "bm25_score": 0.0,
      "combined_score": 0.17175012369064205,
      "rank": 173
    },
    {
      "title": "Manifold learning for fMRI time-varying functional connectivity",
      "authors": [
        "J. Gonzalez-Castillo",
        "Isabel S. Fernandez",
        "K. Lam",
        "D. Handwerker",
        "Francisco Pereira",
        "P. Bandettini"
      ],
      "abstract": "Whole-brain functional connectivity (FC) measured with functional MRI (fMRI) evolves over time in meaningful ways at temporal scales going from years (e.g., development) to seconds [e.g., within-scan time-varying FC (tvFC)]. Yet, our ability to explore tvFC is severely constrained by its large dimensionality (several thousands). To overcome this difficulty, researchers often seek to generate low dimensional representations (e.g., 2D and 3D scatter plots) hoping those will retain important aspects of the data (e.g., relationships to behavior and disease progression). Limited prior empirical work suggests that manifold learning techniques (MLTs)—namely those seeking to infer a low dimensional non-linear surface (i.e., the manifold) where most of the data lies—are good candidates for accomplishing this task. Here we explore this possibility in detail. First, we discuss why one should expect tvFC data to lie on a low dimensional manifold. Second, we estimate what is the intrinsic dimension (ID; i.e., minimum number of latent dimensions) of tvFC data manifolds. Third, we describe the inner workings of three state-of-the-art MLTs: Laplacian Eigenmaps (LEs), T-distributed Stochastic Neighbor Embedding (T-SNE), and Uniform Manifold Approximation and Projection (UMAP). For each method, we empirically evaluate its ability to generate neuro-biologically meaningful representations of tvFC data, as well as their robustness against hyper-parameter selection. Our results show that tvFC data has an ID that ranges between 4 and 26, and that ID varies significantly between rest and task states. We also show how all three methods can effectively capture subject identity and task being performed: UMAP and T-SNE can capture these two levels of detail concurrently, but LE could only capture one at a time. We observed substantial variability in embedding quality across MLTs, and within-MLT as a function of hyper-parameter selection. To help alleviate this issue, we provide heuristics that can inform future studies. Finally, we also demonstrate the importance of feature normalization when combining data across subjects and the role that temporal autocorrelation plays in the application of MLTs to tvFC data. Overall, we conclude that while MLTs can be useful to generate summary views of labeled tvFC data, their application to unlabeled data such as resting-state remains challenging.",
      "url": "https://www.semanticscholar.org/paper/71ae44f01b9201f94779f59aace871f720a77c6d",
      "year": 2023,
      "venue": "Frontiers in Human Neuroscience",
      "source": "semantic_scholar",
      "doi": "10.3389/fnhum.2023.1134012",
      "pdf_url": "https://www.frontiersin.org/articles/10.3389/fnhum.2023.1134012/pdf",
      "citations": 6,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.06911428521217208,
      "novelty_score": 0.9248704663212436,
      "recency_score": 0.7,
      "relevance_score": 0.16313428556365162,
      "bm25_score": 0.0,
      "combined_score": 0.16313428556365162,
      "rank": 174
    },
    {
      "title": "Hyper-holomorphic connections on vector bundles on hyper-Kähler manifolds",
      "authors": [
        "Francesco Meazzini",
        "Claudio Onorati"
      ],
      "abstract": "We study infinitesimal deformations of autodual and hyper-holomorphic connections on complex vector bundles on hyper-Kähler manifolds of arbitrary dimension. In particular, we describe the DG Lie algebra controlling this deformation problem. Moreover, we prove associative formality for derived endomorphisms of a holomorphic vector bundle admitting a projectively hyper-holomorphic connection.",
      "url": "https://www.semanticscholar.org/paper/23584ba7780c416c7064971241be0e1b4fcb2cae",
      "year": 2022,
      "venue": "Mathematische Zeitschrift",
      "source": "semantic_scholar",
      "doi": "10.1007/s00209-022-03176-4",
      "pdf_url": "",
      "citations": 5,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.3039687591901702,
      "novelty_score": 0.9481022463206816,
      "recency_score": 0.6,
      "relevance_score": 0.21319062775705105,
      "bm25_score": 0.0,
      "combined_score": 0.21319062775705105,
      "rank": 175
    },
    {
      "title": "Quantum Substrate Dynamics (QSD): A Relativistic Field Model of Emergent Mass, Inertia and Gravity",
      "authors": [
        "Michael Bush"
      ],
      "abstract": "Quantum Substrate Dynamics (QSD) is a Lorentz-invariant, coherence-based field theory in which mass, gravity, and inertia emerge from phase-stable excitations within a conserved physical substrate. In this framework, mass appears as a coherence-locked phase lattice, inertia arises from reconfiguration resistance at coherence boundaries, and gravity results from large-scale substrate tension gradients. QSD reinterprets black holes, dark matter effects, and cosmological structure as coherence-driven phase transitions within the substrate field—without invoking geometric singularities or exotic matter. While it recovers General Relativity and Quantum Field Theory as limiting cases, QSD also offers falsifiable predictions beyond them, including geometry-sensitive inertia, scalar precursor waves in supernovae, and coherence-based gravitational echoes in black hole mergers. By anchoring known physics in a conserved coherence field, QSD presents a testable and physically unified extension of modern theoretical frameworks.",
      "url": "https://openalex.org/W4411392906",
      "year": 2025,
      "venue": "Preprints.org",
      "source": "openalex",
      "doi": "10.20944/preprints202506.0988.v2",
      "pdf_url": "https://www.preprints.org/frontend/manuscript/724c52c071a67c93de9b5da3679bd484/download_pub",
      "citations": 5,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9280762564991335,
      "recency_score": 0.9,
      "relevance_score": 0.18200000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18200000000000002,
      "rank": 176
    },
    {
      "title": "No Minima, No Collisions: Combining Modulation and Control Barrier Function Strategies for Feasible Dynamical Collision Avoidance",
      "authors": [
        "Yifan Xue",
        "Nadia Figueroa"
      ],
      "abstract": "As prominent real-time safety-critical reactive control techniques, Control Barrier Function Quadratic Programs (CBF-QPs) work for control affine systems in general but result in local minima in the generated trajectories and consequently cannot ensure convergence to the goals. Contrarily, Modulation of Dynamical Systems (Mod-DSs), including normal, reference, and on-manifold Mod-DS, achieve obstacle avoidance with few and even no local minima but have trouble optimally minimizing the difference between the constrained and the unconstrained controller outputs, and its applications are limited to fully-actuated systems. We dive into the theoretical foundations of CBF-QP and Mod-DS, proving that despite their distinct origins, normal Mod-DS is a special case of CBF-QP, and reference Mod-DS's solutions are mathematically connected to that of the CBF-QP through one equation. Building on top of the unveiled theoretical connections between CBF-QP and Mod-DS, reference Mod-based CBF-QP and on-manifold Mod-based CBF-QP controllers are proposed to combine the strength of CBF-QP and Mod-DS approaches and realize local-minimum-free reactive obstacle avoidance for control affine systems in general. We validate our methods in both simulated hospital environments and real-world experiments using Ridgeback for fully-actuated systems and Fetch robots for underactuated systems. Mod-based CBF-QPs outperform CBF-QPs as well as the optimally constrained-enforcing Mod-DS approaches we proposed in all experiments.",
      "url": "https://www.semanticscholar.org/paper/9e252f402a62a45d310efe1391f85e19133e425a",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2502.14238",
      "pdf_url": "",
      "citations": 4,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.035565629288909544,
      "novelty_score": 0.951048951048951,
      "recency_score": 0.9,
      "relevance_score": 0.1922696887866729,
      "bm25_score": 0.0,
      "combined_score": 0.1922696887866729,
      "rank": 177
    },
    {
      "title": "Nonlinear Cauchy Elasticity",
      "authors": [
        "Arash Yavari",
        "Alain Goriely"
      ],
      "abstract": "Abstract Most theories and applications of elasticity rely on an energy function that depends on the strains from which the stresses can be derived. This is the traditional setting of Green elasticity, also known as hyper-elasticity. However, in its original form the theory of elasticity does not assume the existence of a strain energy function. In this case, called Cauchy elasticity, stresses are directly related to the strains. Since the emergence of modern elasticity in the 1940s, research on Cauchy elasticity has been relatively limited. One possible reason for this is that for Cauchy materials, the net work performed by stress along a closed path in the strain space may be nonzero. Therefore, such materials may require access to both energy sources and sinks. This characteristic has led some mechanicians to question the viability of Cauchy elasticity as a physically plausible theory of elasticity. In this paper, motivated by its relevance to recent applications, such as the modeling of active solids, we revisit Cauchy elasticity in a modern form. First, we show that in the general theory of anisotropic Cauchy elasticity, stress can be expressed in terms of six functions, that we call Edelen-Darboux potentials . For isotropic Cauchy materials, this number reduces to three, while for incompressible isotropic Cauchy elasticity, only two such potentials are required. Second, we show that in Cauchy elasticity, the link between balance laws and symmetries is lost, in general, ",
      "url": "https://openalex.org/W4413891066",
      "year": 2025,
      "venue": "Archive for Rational Mechanics and Analysis",
      "source": "openalex",
      "doi": "10.1007/s00205-025-02120-0",
      "pdf_url": "https://link.springer.com/content/pdf/10.1007/s00205-025-02120-0.pdf",
      "citations": 4,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.013199871692412423,
      "novelty_score": 0.9870967741935485,
      "recency_score": 0.9,
      "relevance_score": 0.18555996150772375,
      "bm25_score": 0.0,
      "combined_score": 0.18555996150772375,
      "rank": 178
    },
    {
      "title": "Mock Modularity at Work, or Black Holes in a Forest",
      "authors": [
        "Sergei Alexandrov"
      ],
      "abstract": "Mock modular forms, first invented by Ramanujan, provide a beautiful generalization of the usual modular forms. In recent years, it was found that they capture the generating functions of the number of microstates of BPS black holes appearing in compactifications of string theory with 8 and 16 supercharges. This review describes these results and their applications, which range from the actual computation of these generating functions for both compact and non-compact compactification manifolds (encoding, respectively, Donaldson–Thomas and Vafa–Witten topological invariants) to the construction of new non-commutative structures on moduli spaces of Calabi–Yau threefolds.",
      "url": "https://openalex.org/W4411972976",
      "year": 2025,
      "venue": "Entropy",
      "source": "openalex",
      "doi": "10.3390/e27070719",
      "pdf_url": "https://www.mdpi.com/1099-4300/27/7/719/pdf?version=1751469528",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9532710280373832,
      "recency_score": 0.9,
      "relevance_score": 0.18120000000000003,
      "bm25_score": 0.0,
      "combined_score": 0.18120000000000003,
      "rank": 179
    },
    {
      "title": "Estimating Functional Brain Networks by Low-Rank Representation With Local Constraint",
      "authors": [
        "Zhigang Li",
        "Weimin Zheng",
        "Honghong Liu",
        "Jingyu Liu",
        "Chang Yan",
        "Zhiqun Wang",
        "Bin Hu",
        "Qunxi Dong"
      ],
      "abstract": "The functional architecture undergoes alterations during the preclinical phase of Alzheimer’s disease. Consequently, the primary research focus has shifted towards identifying Alzheimer’s disease and its early stages by constructing a functional connectivity network based on resting-state fMRI data. Recent investigations show that as Alzheimer’s Disease (AD) progresses, modular tissue and connections in the core brain areas of AD patients diminish. Sparse learning methods are powerful tools for understanding Functional Brain Networks (FBNs) with Regions of Interest (ROIs) and a connectivity matrix measuring functional coherence between them. However, these tools often focus exclusively on functional connectivity measures, neglecting the brain network’s modularity. Modularity orchestrates dynamic activities within the FBN to execute intricate cognitive tasks. To provide a comprehensive delineation of the FBN, we propose a local similarity-constrained low-rank sparse representation (LSLRSR) method that encodes modularity information under a manifold-regularized network learning framework and further formulate it as a low-rank sparse graph learning problem, which can be solved by an efficient optimization algorithm. Specifically, for each modularity structure, the Schatten p-norm regularizer reduces the reconstruction error and provides a better approximation of the low-rank constraint. Furthermore, we adopt a manifold-regularized local similarity prior to infer the intricate relationship between subnetwork similarity and modularity, guiding the modeling of FBN. Additionally, the proximal average method approximates the joint solution’s proximal map, and the resulting nonconvex optimization problems are solved using the alternating direction multiplier method (ADMM). Compared to state-of-the-art methods for constructing FBNs, our algorithm generates a more modular FBN. This lays the groundwork for further research into alterations in brain network modularity resulting from diseases.",
      "url": "https://www.semanticscholar.org/paper/8f0d342861da43a0c039cbb16c33bfab55b54d5b",
      "year": 2024,
      "venue": "IEEE transactions on neural systems and rehabilitation engineering",
      "source": "semantic_scholar",
      "doi": "10.1109/TNSRE.2024.3355769",
      "pdf_url": "https://ieeexplore.ieee.org/ielx7/7333/4359219/10403846.pdf",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0417384228859795,
      "novelty_score": 0.9695817490494297,
      "recency_score": 0.8,
      "relevance_score": 0.1737215268657939,
      "bm25_score": 0.0,
      "combined_score": 0.1737215268657939,
      "rank": 180
    },
    {
      "title": "Clarke Transform and Clarke Coordinates - A New Kid on the Block for State Representation of Continuum Robots",
      "authors": [
        "R. Grassmann",
        "J. Burgner-Kahrs"
      ],
      "abstract": "For almost all tendon-driven continuum robots, a segment is actuated by three or four tendons constrained by its mechanical design. For both cases, methods to account for the constraints are known. However, for an arbitrary number of tendons, a disentanglement method has yet to be formulated. Motivated by this unsolved general case, we explored state representations and exploited the two-dimensional manifold. We found that the Clarke transformation, a mathematical transformation used in vector control, can be generalized to address this problem. We present the Clarke transform and Clarke coordinates, which can be used to overcome the troublesome interdependency between the tendons, simplify modeling, and unify different improved state representations. Further connection to arc parameters leads to the possibility to derive more generalizable approaches applicable to a wider range of robot types.",
      "url": "https://www.semanticscholar.org/paper/b085d64d385554afe5b567875494a232cb6f9f9f",
      "year": 2024,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2409.13826",
      "pdf_url": "",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.02220840242184171,
      "novelty_score": 0.9119205298013244,
      "recency_score": 0.8,
      "relevance_score": 0.16786252072655256,
      "bm25_score": 0.0,
      "combined_score": 0.16786252072655256,
      "rank": 181
    },
    {
      "title": "A Lightweight Certificateless Signcryption Scheme based on HCC for securing Underwater Wireless Sensor Networks (UWSNs)",
      "authors": [
        "Meenakshi Gupta",
        "Poonam Gera",
        "Bharavi Mishra"
      ],
      "abstract": "Underwater Wireless Sensor Networks (UWSNs) consist of sensor nodes deployed within bodies of water. Wireless connections and the harsh underwater environment make sensors susceptible to a variety of malevolent attacks and security concerns. The fundamental concern of the UWSN is secure and reliable communication with low energy consumption. Many cryptographic solutions have been proposed to deal with such constraints. Signcryption is one of these cryptosystems, integrating signature and encryption to reduce computational costs relative to other cryptosystems. Numerous signcryption schemes based on ElGamal, bilinear pairing, RSA, and Elliptic Curve Cryptography (ECC) have been proposed. The inadequacies of these schemes include increased computation and communication overhead, the absence of certain security features, and a high memory demand. In this study, we introduced a lightweight certificateless signcryption system based on Hyper-elliptic Curve Cryptography (HCC). The system significantly reduces computing and communication costs, making it ideal for resource-constrained environments. Our solution meets the necessary security requirements while maintaining forward secrecy. The security analysis of our scheme indicates that the proposed method is efficient and effective.",
      "url": "https://www.semanticscholar.org/paper/8dfa770c6dfc24568c6ae831e0e133a97b3c0c25",
      "year": 2023,
      "venue": "International Conference on Security of Information and Networks",
      "source": "semantic_scholar",
      "doi": "10.1109/SIN60469.2023.10474770",
      "pdf_url": "",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.05816958761795378,
      "novelty_score": 0.9460016488046167,
      "recency_score": 0.7,
      "relevance_score": 0.15865087628538613,
      "bm25_score": 0.0,
      "combined_score": 0.15865087628538613,
      "rank": 182
    },
    {
      "title": "Cognitive Computing Continuum: State-of-the-Art Review and ENACT Vision &amp; Approach",
      "authors": [
        "Ioanna Angeliki Kapetanidou",
        "Alexandros Nizamis",
        "Efstathios Karanastasis",
        "Gabriel-Mihail Danciu",
        "Clara Isabel Valero López",
        "Thanasis Kotsiopoulos",
        "Jordi Gascón",
        "Jaime Flor",
        "Ross Campbell",
        "Athanasios Liatifis"
      ],
      "abstract": "Abstract The evolution from the Edge-Cloud Continuum to the Cognitive Computing Continuum (CCC) has introduced new challenges which necessitate advanced frameworks that integrate cognitive capabilities to enhance interoperability, adaptability, and resource efficiency. Considering insights from ongoing research and initiatives on the cognitive cloud, we identify the core concepts essential for transitioning to the CCC, including cognitive orchestration, distributed AI, and sovereign data management. We then introduce ENACT, a novel framework designed to embrace these concepts aiming to provide cognitive, highly adaptive orchestration to support modern hyper-distributed and data-intensive applications. Furthermore, ENACT employs bespoke mechanisms to enable dynamic continuum modelling and visibility as well as to facilitate application-level automation and adaptability. This paper presents the motivation behind the ENACT approach, reviews the state-of-the-art across its fundamental technological concepts and highlights its key innovations. Overall, it contributes to the formalization of the CCC architectural principles and, ultimately, to the realization of CCC.",
      "url": "https://openalex.org/W4413129303",
      "year": 2025,
      "venue": "Journal of Grid Computing",
      "source": "openalex",
      "doi": "10.1007/s10723-025-09810-9",
      "pdf_url": "https://link.springer.com/content/pdf/10.1007/s10723-025-09810-9.pdf",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.022332340534288224,
      "novelty_score": 0.959849435382685,
      "recency_score": 0.9,
      "relevance_score": 0.1874997021602865,
      "bm25_score": 0.0,
      "combined_score": 0.1874997021602865,
      "rank": 183
    },
    {
      "title": "Kalb-Ramond field, black holes and black strings in (2 + 1)D",
      "authors": [
        "Meseret Asrat"
      ],
      "abstract": "A bstract New rotating dilaton black hole and black string solutions in three spacetime dimensions are obtained. The solutions are asymptotically flat and, they are exact in classical string theory. The black hole solutions have only a single horizon. Enclosed inside their horizons, they contain a curvature singularity. The black strings carry axion charges and have two horizons. Depending on the ratio of their inner and outer horizons radii, they may or may not contain a curvature singularity. When they contain a singularity, the singularity is either at or enclosed inside their inner horizons. We also show that a solution with a constant or asymptotically non-zero Kalb-Ramond field is equivalent to a solution with no Kalb-Ramond field and non-zero rotation at asymptotic infinity. In general, non-zero rotation allows negative mass solutions. To demonstrate, we give negative mass black hole solutions. We also discuss black hole and black string solutions with a curvature singularity at or beyond their outer or event horizons. We also present novel black hole solutions with a ring curvature singularity in between their inner and outer horizons.",
      "url": "https://openalex.org/W4413471887",
      "year": 2025,
      "venue": "Journal of High Energy Physics",
      "source": "openalex",
      "doi": "10.1007/jhep08(2025)135",
      "pdf_url": "https://link.springer.com/content/pdf/10.1007/JHEP08(2025)135.pdf",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9465691788526434,
      "recency_score": 0.9,
      "relevance_score": 0.18080000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18080000000000002,
      "rank": 184
    },
    {
      "title": "Codepoietic Generation of Meaningful Information in the Evolving Biosphere",
      "authors": [
        "Abir U. Igamberdiev"
      ],
      "abstract": "Meaningful information represents reality in its potential form, and its actualization increases the system’s negentropy. Biological evolution leads to the expansion of meaningful information by generating new coding systems (codepoiesis). Through this expansion, any evolutionary change obtains functional value when it receives an interpretation through which it gives rise to a meaningful function. Complexification in the evolutionary process corresponds to the generation of new meaningful information and, thus, to the development of new structures with corresponding functions. Any biological function has a meaning within the context of a particular environment, and the evolutionary search for new meanings results in the establishment of the state of sustainable non-equilibrium acting as an attractor, in which the developing system achieves the condition of maximization of its power via synergistic effects. At higher levels of the organization, evolutionary innovations emerge as niche constructions, behavioral choices, and, finally, the phenomenon of cognition. The evolutionary growth of meanings appears as a part of the expanding information system formed by the organisms inhabiting it. It acquires major expansion with the emergence of consciousness that incorporates the image of the whole world into the dynamic process of knowledge acquisition and creates the conditions for the development of global civilization.",
      "url": "https://openalex.org/W4411607412",
      "year": 2025,
      "venue": "Entropy",
      "source": "openalex",
      "doi": "10.3390/e27070672",
      "pdf_url": "https://www.mdpi.com/1099-4300/27/7/672/pdf?version=1750738707",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9198396793587174,
      "recency_score": 0.9,
      "relevance_score": 0.18080000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18080000000000002,
      "rank": 185
    },
    {
      "title": "Coriolis Factorizations and their Connections to Riemannian Geometry",
      "authors": [
        "Patrick M. Wensing",
        "Johannes Englsberger",
        "Jean-Jacques E. Slotine"
      ],
      "abstract": "Many energy-based control strategies for mechanical systems require the choice of a Coriolis factorization satisfying a skew-symmetry property. This paper (a) explores if and when a control designer has flexibility in this choice, (b) develops a canonical choice related to the Christoffel symbols, and (c) describes how to efficiently perform control computations with it for constrained mechanical systems. We link the choice of a Coriolis factorization to the notion of an affine connection on the configuration manifold and show how properties of the connection relate with the associated factorization. In particular, the factorization based on the Christoffel symbols is linked with a torsion-free property that can limit the twisting of system trajectories during passivity-based control. We then develop a way to induce Coriolis factorizations for constrained mechanisms from unconstrained ones, which provides a pathway to use the theory for efficient control computations with high-dimensional systems such as humanoids and quadruped robots with open- and closed-chain mechanisms. A collection of algorithms is provided (and made available open source) to support the recursive computation of passivity-based control laws, adaptation laws, and regressor matrices in future applications.",
      "url": "https://www.semanticscholar.org/paper/4aeccd0f9f21f06a686422b3ef0c52455ca48a20",
      "year": 2023,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2312.14425",
      "pdf_url": "",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.057598877157715514,
      "novelty_score": 0.9422632794457275,
      "recency_score": 0.7,
      "relevance_score": 0.15807966314731464,
      "bm25_score": 0.0,
      "combined_score": 0.15807966314731464,
      "rank": 186
    },
    {
      "title": "Hyper-graph regularized subspace clustering with skip connections for band selection of hyperspectral image",
      "authors": [
        "Meng Zeng",
        "Bin Ning",
        "Qiong Gu",
        "Chunyang Hu",
        "Shuijia Li"
      ],
      "abstract": "The Hughes phenomenon of Hyperspectral images (HSIs) with the hundreds of\n continuous narrow bands makes the computational cost of HSIs process ing\n high. Band selection is an effective way to solve such a problem and a lot\n of band selection methods have been proposed in recent years. In this paper,\n a novel hyper-graph regularized subspace clustering with skip connections\n (HRSC-SC) is proposed for band selection of hyperspectral image, which is a\n clustering-based band selection method. The networks combine subspace\n clustering into the convolutional auto-encoder by thinking of it as a\n self-expressive layer. To make full use of the historical feature maps\n obtained from the networks and tackle the problem of gradient vanishing\n caused by multiple nonlinear transformations, the symmetrical skip\n connections are added to the networks to pass image details from encoder to\n decoder. Furthermore, the hyper-graph regularization is presented to\n consider the manifold structure reflecting geometric information within\n data, which accurately describes the multivariate relationship between data\n points and makes the results of clustering more accurate so that select the\n most representative band subset. The proposed HRSC-SC band selection method\n is compared with the existing robust band selection algorithms on Indian\n Pines, Salinas-A, and Pavia University HSIs, showing that the results of\n the proposed method outperform the current state-of-the-art band selection\n methods. Especially, the overall accuracy of the clustering is the best on\n three real HSIs compared to other methods when the band selection number is\n 25, reaching 82.62%, 92.48%, and 96,5% respectively.",
      "url": "https://www.semanticscholar.org/paper/280a90ebc45e900fd837290bd7f868c37dc04f9c",
      "year": 2022,
      "venue": "Computer Science and Information Systems",
      "source": "semantic_scholar",
      "doi": "10.2298/csis210830005z",
      "pdf_url": "http://www.doiserbia.nb.rs/ft.aspx?id=1820-02142200005Z",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.10970544454460424,
      "novelty_score": 0.9294392523364486,
      "recency_score": 0.6,
      "relevance_score": 0.15371163336338126,
      "bm25_score": 0.0,
      "combined_score": 0.15371163336338126,
      "rank": 187
    },
    {
      "title": "Regularity of CR maps into uniformly pseudo convex hyper surfaces and applications to proper holomorphic maps",
      "authors": [
        "Josef Greilhuber",
        "B. Lamel"
      ],
      "abstract": "We study regularity properties of CR maps in positive codimension valued in pseudoconvex manifolds which carry a nontrivial Levi foliation. We introduce an invariant which can be used to deduce that any sufficiently regular CR map from a minimal manifold into such a foliated target is either generically smooth or geometrically highly constrained, and to show generic smoothness of sufficiently regular CR transversal CR maps between pseudoconvex hypersurfaces. As an application, we discuss boundary regularity of proper holomorphic maps into bounded symmetric domains.",
      "url": "https://www.semanticscholar.org/paper/076fb2e5f9317a8551e59d2beb7a1c0939f083e4",
      "year": 2022,
      "venue": "ANNALI SCUOLA NORMALE SUPERIORE - CLASSE DI SCIENZE",
      "source": "semantic_scholar",
      "doi": "10.2422/2036-2145.202105_009",
      "pdf_url": "https://arxiv.org/pdf/2302.14016",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.04899204886492442,
      "novelty_score": 0.9536423841059603,
      "recency_score": 0.6,
      "relevance_score": 0.13549761465947732,
      "bm25_score": 0.0,
      "combined_score": 0.13549761465947732,
      "rank": 188
    },
    {
      "title": "Cross-Dimensional Mathematics: A Foundation For STP/STA",
      "authors": [
        "Daizhan Cheng"
      ],
      "abstract": "A new mathematical structure, called the cross-dimensional mathematics (CDM), is proposed. The CDM considered in this paper consists of three parts: hyper algebra, hyper geometry, and hyper Lie group/Lie algebra. Hyper algebra proposes some new algebraic structures such as hyper group, hyper ring, and hyper module over matrices and vectors with mixed dimensions (MVMDs). They have sets of classical groups, rings, and modules as their components and cross-dimensional connections among their components. Their basic properties are investigated. Hyper geometry starts from mixed dimensional Euclidian space, and hyper vector space. Then the hyper topological vector space, hyper inner product space, and hyper manifold are constructed. They have a joined cross-dimensional geometric structure. Finally, hyper metric space, topological hyper group and hyper Lie algebra are built gradually, and finally, the corresponding hyper Lie group is introduced. All these concepts are built over MVMDs, and to reach our purpose in addition to existing semi-tensor products (STPs) and semi-tensor additions (STAs), a couple of most general STP and STA are introduced. Some existing structures/results about STPs/STAs have also been resumed and integrated into this CDM.",
      "url": "https://www.semanticscholar.org/paper/39270d175ed49f35e036ca21ee61e7490c5d627f",
      "year": 2024,
      "venue": "",
      "source": "semantic_scholar",
      "doi": "10.1007/s11425-024-2528-4",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.3397085375332654,
      "novelty_score": 0.9080118694362017,
      "recency_score": 0.8,
      "relevance_score": 0.26231256125997965,
      "bm25_score": 0.0,
      "combined_score": 0.26231256125997965,
      "rank": 189
    },
    {
      "title": "Manifold Trajectories in Next-Token Prediction: From Replicator Dynamics to Softmax Equilibrium",
      "authors": [
        "Christopher R. Lee-Jenkins"
      ],
      "abstract": "Decoding in large language models is often described as scoring tokens and normalizing with softmax. We give a minimal, self-contained account of this step as a constrained variational principle on the probability simplex. The discrete, normalization-respecting ascent is the classical multiplicative-weights (entropic mirror) update; its continuous-time limit is the replicator flow. From these ingredients we prove that, for a fixed context and temperature, the next-token distribution follows a smooth trajectory inside the simplex and converges to the softmax equilibrium. This formalizes the common ``manifold traversal''intuition at the output-distribution level. The analysis yields precise, practice-facing consequences: temperature acts as an exact rescaling of time along the same trajectory, while top-k and nucleus sampling restrict the flow to a face with identical guarantees. We also outline a controlled account of path-dependent score adjustments and their connection to loop-like, hallucination-style behavior. We make no claims about training dynamics or internal representations; those are deferred to future work.",
      "url": "https://www.semanticscholar.org/paper/5b675c8e73259f41c0ed2b9ab4dad68f21111dc2",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2508.21186",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.032278450905079824,
      "novelty_score": 0.9639175257731959,
      "recency_score": 0.9,
      "relevance_score": 0.19008353527152397,
      "bm25_score": 0.0,
      "combined_score": 0.19008353527152397,
      "rank": 190
    },
    {
      "title": "Federated Learning for Large-Scale Cloud Robotic Manipulation: Opportunities and Challenges",
      "authors": [
        "Obaidullah Zaland",
        "Chanh Nguyen",
        "Florian T. Pokorny",
        "Monowar H. Bhuyan"
      ],
      "abstract": "Federated Learning (FL) is an emerging distributed machine learning paradigm, where the collaborative training of a model involves dynamic participation of devices to achieve broad objectives. In contrast, classical machine learning (ML) typically requires data to be located on-premises for training, whereas FL leverages numerous user devices to train a shared global model without the need to share private data. Current robotic manipulation tasks are constrained by the individual capabilities and speed of robots due to limited low-latency computing resources. Consequently, the concept of cloud robotics has emerged, allowing robotic applications to harness the flexibility and reliability of computing resources, effectively alleviating their computational demands across the cloud-edge continuum. Undoubtedly, within this distributed computing context, as exemplified in cloud robotic manipulation scenarios, FL offers manifold advantages while also presenting several challenges and opportunities. In this paper, we present fundamental concepts of FL and their connection to cloud robotic manipulation. Additionally, we envision the opportunities and challenges associated with realizing efficient and reliable cloud robotic manipulation at scale through FL, where researchers adapt to design and verify FL models in either centralized or decentralized settings.",
      "url": "https://www.semanticscholar.org/paper/810506e46071c6581f9c2f608e584339fc713926",
      "year": 2025,
      "venue": "International Conference on Machine Learning and Computing",
      "source": "semantic_scholar",
      "doi": "10.1109/ICMLC66258.2025.11280176",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.014945488101438066,
      "novelty_score": 0.922242314647378,
      "recency_score": 0.9,
      "relevance_score": 0.18488364643043145,
      "bm25_score": 0.0,
      "combined_score": 0.18488364643043145,
      "rank": 191
    },
    {
      "title": "A Connection Between Score Matching and Local Intrinsic Dimension",
      "authors": [
        "Eric Yeats",
        "Aaron Jacobson",
        "Darryl Hannan",
        "Yiran Jia",
        "Timothy Doster",
        "H. Kvinge",
        "Scott Mahan"
      ],
      "abstract": "The local intrinsic dimension (LID) of data is a fundamental quantity in signal processing and learning theory, but quantifying the LID of high-dimensional, complex data has been a historically challenging task. Recent works have discovered that diffusion models capture the LID of data through the spectra of their score estimates and through the rate of change of their density estimates under various noise perturbations. While these methods can accurately quantify LID, they require either many forward passes of the diffusion model or use of gradient computation, limiting their applicability in compute- and memory-constrained scenarios. We show that the LID is a lower bound on the denoising score matching loss, motivating use of the denoising score matching loss as a LID estimator. Moreover, we show that the equivalent implicit score matching loss also approximates LID via the normal dimension and is closely related to a recent LID estimator, FLIPD. Our experiments on a manifold benchmark and with Stable Diffusion 3.5 indicate that the denoising score matching loss is a highly competitive and scalable LID estimator, achieving superior accuracy and memory footprint under increasing problem size and quantization level.",
      "url": "https://www.semanticscholar.org/paper/6bcd7f677e1dbfc1463f4c1762e3a278fd2d692f",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2510.12975",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.014469235608870425,
      "novelty_score": 0.9348268839103869,
      "recency_score": 0.9,
      "relevance_score": 0.18474077068266115,
      "bm25_score": 0.0,
      "combined_score": 0.18474077068266115,
      "rank": 192
    },
    {
      "title": "Research on Time Series Prediction Model of Quantum Long Short Term Memory Network Fusion",
      "authors": [
        "Bing Han",
        "Jian Kang",
        "Hongyu Su"
      ],
      "abstract": "This study proposes a novel hybrid prediction model (QGCN-LSTM) that combines quantum graph convolutional networks with classical LSTM. The model takes classical time series data as input and achieves classical quantum information conversion through a quantum encoding layer. Multi scale features are extracted through the collaborative computation of quantum graph convolutional modules (QGCN) and quantum gated loop units, and a quantum attention module is introduced to dynamically screen key information. Finally, the prediction results are generated through quantum measurement and classical output layer. In the time series prediction task of urban traffic flow, a benchmark model system covering classical, cutting-edge, and traditional architectures was constructed. The experimental results show that QGCN-LSTM utilizes quantum entanglement gates to establish non local road network associations, dynamically improves key node weights based on quantum state fidelity, and achieves deep compression of lines through quantum line pruning technology, effectively alleviating the common problem of &amp;quot;poor plateau&amp;quot; in quantum neural network training. In terms of prediction accuracy, the average absolute error (MAE) of its key hub nodes is reduced by 34.1% compared to the classical graph convolution LSTM (GCN-LSTM) model, and the spatial correlation index (SCI) is improved to 0.89. In addition, it also shows excellent performance in dynamic response, edge computing efficien",
      "url": "https://openalex.org/W4413114546",
      "year": 2025,
      "venue": "Preprints.org",
      "source": "openalex",
      "doi": "10.20944/preprints202508.0647.v1",
      "pdf_url": "https://www.preprints.org/frontend/manuscript/cd23a06aae90a6a104783f136fdc0cc1/download_pub",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9507323568575232,
      "recency_score": 0.9,
      "relevance_score": 0.18040000000000003,
      "bm25_score": 0.0,
      "combined_score": 0.18040000000000003,
      "rank": 193
    },
    {
      "title": "Neural Architecture Search for Hyperspectral Image Classification: A Comprehensive Review and Future Perspectives",
      "authors": [
        "Aili Wang",
        "Xinyu Liu",
        "Kang Zhang",
        "Haoran Lv",
        "Haibin Wu",
        "Xing Chen",
        "Mudi Yao"
      ],
      "abstract": "Hyperspectral image classification (HSIC) is a key task in the field of remote sensing, but the complex nature of hyperspectral data poses a serious challenge to traditional methods. Although deep learning significantly improves classification performance through automatic feature extraction, manually designed network architectures suffer from issues such as dependence on expert experience and lack of flexibility. Neural architecture search (NAS) provides new ideas for HSIC through automated network structure optimization. This article systematically reviews the application progress of NAS in HSIC: firstly, the core components of NAS are analyzed, and the characteristics of various methods are compared from three aspects: search space, search strategy, and performance evaluation. Furthermore, the focus is on exploring NAS technology based on convolutional neural networks, covering 1D, 2D, and 3D convolutional architectures and their innovative integration with various technologies, revealing the advantages of NAS in HSIC. However, NAS still faces challenges such as high computing resource requirements and insufficient interpretability. This article systematically reviews the application of NAS in the field of HSIC for the first time, facilitating readers to quickly understand the development process of NAS in HSIC and the advantages and disadvantages of various technologies, proposing possible future research directions.",
      "url": "https://openalex.org/W4413037406",
      "year": 2025,
      "venue": "Remote Sensing",
      "source": "openalex",
      "doi": "10.3390/rs17152727",
      "pdf_url": "https://www.mdpi.com/2072-4292/17/15/2727/pdf?version=1754553071",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9221140472878999,
      "recency_score": 0.9,
      "relevance_score": 0.18040000000000003,
      "bm25_score": 0.0,
      "combined_score": 0.18040000000000003,
      "rank": 194
    },
    {
      "title": "Microstates of AdS5 black holes with hypermultiplets",
      "authors": [
        "Marina David",
        "Annelien Vekemans"
      ],
      "abstract": "A bstract We construct supersymmetric rotating AdS 5 black holes in 5d $$ \\mathcal{N} $$ <mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"> <mml:mi>N</mml:mi> </mml:math> = 2 gauged supergravity coupled to two vector multiplets and a universal hypermultiplet, and verify their microscopic counting from the superconformal index of the dual 4d class $$ \\mathcal{S} $$ <mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"> <mml:mi>S</mml:mi> </mml:math> $$ \\mathcal{N} $$ <mml:math xmlns:mml=\"http://www.w3.org/1998/Math/MathML\"> <mml:mi>N</mml:mi> </mml:math> = 1 SCFTs. From the CFT, we perform the Legendre transform of the index to the microcanonical ensemble. The theories are parametrized by a rational number z which enters into the extremization equations making them more challenging to solve. We present a method to address these difficulties and highlight the subtleties involved. From the gravity perspective, we identify a charged, rotating black hole whose Bekenstein-Hawking entropy matches the prediction from the index for z = ±1. Beyond this value, where hypermultiplet scalars are nonzero, we construct the near-horizon extremal geometry perturbatively around z = 1 and verify that the entropy is consistent with the CFT prediction. We discuss the thermodynamics and verify the near-horizon versions of the first law of thermodynamics and the supersymmetric condition. In this setting, our analysis characterizes the first construction of a rotating black hole geometry in",
      "url": "https://openalex.org/W4412403714",
      "year": 2025,
      "venue": "Journal of High Energy Physics",
      "source": "openalex",
      "doi": "10.1007/jhep07(2025)148",
      "pdf_url": "https://link.springer.com/content/pdf/10.1007/JHEP07(2025)148.pdf",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9288811795316566,
      "recency_score": 0.9,
      "relevance_score": 0.18040000000000003,
      "bm25_score": 0.0,
      "combined_score": 0.18040000000000003,
      "rank": 195
    },
    {
      "title": "A Unified 4D Quantum Projection Framework of Space, Time, and Measurement",
      "authors": [
        "Mazen Zaino"
      ],
      "abstract": "This paper presents a novel theoretical framework that aims to unify the core principles of quantum mechanics, general relativity, and thermodynamics by introducing an extended spatial geometry incorporating a compactified fourth spatial dimension. The theory proposes that many of the counterintuitive behaviours observed in quantum systems, such as wave function collapse, superposition, entanglement, tunnelling, and decoherence, can be naturally explained as consequences of 4D quantum objects being projected into 3D space. In this framework, observable quantum behaviour emerges not from intrinsic randomness but from the limitations of a lower-dimensional perspective on higher-dimensional structures. At the heart of this unification is the introduction of a new scalar field, termed the entropion field, which governs the rate and nature of quantum decoherence while simultaneously encoding the thermodynamic arrow of time. The entropion field interacts with known quantum fields and gravitational curvature, modifying the Einstein field equations and extending the standard model of quantum field theory. The result is a coherent description of how classical reality emerges from quantum substrates, how entropy and time are fundamentally linked, and how spacetime geometry influences quantum behaviour. The paper develops this hypothesis through a structured hierarchy of conceptual foundations, mathematical formalism, physical interpretation, and predictive consequences. Particular atte",
      "url": "https://openalex.org/W4411509857",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.14293/pr2199.001746.v1",
      "pdf_url": "https://www.scienceopen.com/document_file/6cb529af-1a5b-4d7d-9a3e-114650db0c6c/ScienceOpenPreprint/A%20Unified%204D%20Quantum%20Projection%20Framework.pdf",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9141770776751765,
      "recency_score": 0.9,
      "relevance_score": 0.18040000000000003,
      "bm25_score": 0.0,
      "combined_score": 0.18040000000000003,
      "rank": 196
    },
    {
      "title": "A Survey of Task-Oriented Knowledge Graph Reasoning: Status, Applications, and Prospects",
      "authors": [
        "Guanglin Niu",
        "Bo Li",
        "Yangguang Lin"
      ],
      "abstract": "",
      "url": "https://openalex.org/W4411222678",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.36227/techrxiv.174961563.32605293/v1",
      "pdf_url": "https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.174961563.32605293/v1",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9226973684210527,
      "recency_score": 0.9,
      "relevance_score": 0.18040000000000003,
      "bm25_score": 0.0,
      "combined_score": 0.18040000000000003,
      "rank": 197
    },
    {
      "title": "CiliaGraph: Enabling Expression-enhanced Hyper-Dimensional Computation in Ultra-Lightweight and One-Shot Graph Classification on Edge",
      "authors": [
        "Yuxi Han",
        "Jihe Wang",
        "Danghui Wang"
      ],
      "abstract": "Graph Neural Networks (GNNs) are computationally demanding and inefficient when applied to graph classification tasks in resource-constrained edge scenarios due to their inherent process, involving multiple rounds of forward and backward propagation. As a lightweight alternative, Hyper-Dimensional Computing (HDC), which leverages high-dimensional vectors for data encoding and processing, offers a more efficient solution by addressing computational bottleneck. However, current HDC methods primarily focus on static graphs and neglect to effectively capture node attributes and structural information, which leads to poor accuracy. In this work, we propose CiliaGraph, an enhanced expressive yet ultra-lightweight HDC model for graph classification. This model introduces a novel node encoding strategy that preserves relative distance isomorphism for accurate node connection representation. In addition, node distances are utilized as edge weights for information aggregation, and the encoded node attributes and structural information are concatenated to obtain a comprehensive graph representation. Furthermore, we explore the relationship between orthogonality and dimensionality to reduce the dimensions, thereby further enhancing computational efficiency. Compared to the SOTA GNNs, extensive experiments show that CiliaGraph reduces memory usage and accelerates training speed by an average of 292 times(up to 2341 times) and 103 times(up to 313 times) respectively while maintaining comparable accuracy.",
      "url": "https://www.semanticscholar.org/paper/f5d6c679fcc0d9149abe51b9bea8905072593290",
      "year": 2024,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2405.19033",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.050036463946114655,
      "novelty_score": 0.9435483870967741,
      "recency_score": 0.8,
      "relevance_score": 0.17541093918383444,
      "bm25_score": 0.0,
      "combined_score": 0.17541093918383444,
      "rank": 198
    },
    {
      "title": "On the selection of a proper connection in describing the dynamics of constrained mechanical systems",
      "authors": [
        "S. Natsiavas",
        "P. Passas",
        "K. Tzaferis"
      ],
      "abstract": "This study is focused on a critical issue related to the direct and consistent application of Newton’s law of motion to a special but large class of mechanical systems, involving equality motion constraints. For these systems, it is advantageous to employ the general analytical dynamics framework, where their motion is represented by a curve on a non-flat configuration manifold. The geometric properties of this manifold, providing the information needed for setting up the equations of motion of the system examined, are fully determined by two mathematical entities. The first of them is the metric tensor, whose components at each point of the manifold are obtained by considering the kinetic energy of the system. The second geometric entity is known as the connection of the manifold. In dynamics, the components of the connection are established by using the set of the motion constraints imposed on the original system and provide the torsion and curvature properties of the manifold. Despite its critical role in the dynamics of constrained systems, the significance of the connection has not been investigated yet at a sufficient level in the current engineering literature. The main objective of the present work is to first provide a systematic way for selecting this geometric entity and then illustrate its role and importance in describing the dynamics of constrained systems.",
      "url": "https://www.semanticscholar.org/paper/a95c3704bec9d3c60ba0bf06ed6c9cc2d8d10392",
      "year": 2023,
      "venue": "Nonlinear dynamics",
      "source": "semantic_scholar",
      "doi": "10.1007/s11071-023-08326-9",
      "pdf_url": "https://link.springer.com/content/pdf/10.1007/s11071-023-08326-9.pdf",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.08373903974804225,
      "novelty_score": 0.8898797983714619,
      "recency_score": 0.7,
      "relevance_score": 0.16552171192441267,
      "bm25_score": 0.0,
      "combined_score": 0.16552171192441267,
      "rank": 199
    },
    {
      "title": "An efficient constraint method for solving planning problems under end-effector constraints",
      "authors": [
        "Yahao Wang",
        "Zhen Li",
        "Yanghong Li",
        "Erbao Dong"
      ],
      "abstract": "\nPurpose\nIn response to the challenge of reduced efficiency or failure of robot motion planning algorithms when faced with end-effector constraints, this study aims to propose a new constraint method to improve the performance of the sampling-based planner.\n\n\nDesign/methodology/approach\nIn this work, a constraint method (TC method) based on the idea of cross-sampling is proposed. This method uses the tangent space in the workspace to approximate the constrained manifold pattern and projects the entire sampling process into the workspace for constraint correction. This method avoids the need for extensive computational work involving multiple iterations of the Jacobi inverse matrix in the configuration space and retains the sampling properties of the sampling-based algorithm.\n\n\nFindings\nSimulation results demonstrate that the performance of the planner when using the TC method under the end-effector constraint surpasses that of other methods. Physical experiments further confirm that the TC-Planner does not cause excessive constraint errors that might lead to task failure. Moreover, field tests conducted on robots underscore the effectiveness of the TC-Planner, and its excellent performance, thereby advancing the autonomy of robots in power-line connection tasks.\n\n\nOriginality/value\nThis paper proposes a new constraint method combined with the rapid-exploring random trees algorithm to generate collision-free trajectories that satisfy the constraints for a high-dimensional robotic system under end-effector constraints. In a series of simulation and experimental tests, the planner using the TC method under end-effector constraints efficiently performs. Tests on a power distribution live-line operation robot also show that the TC method can greatly aid the robot in completing operation tasks with end-effector constraints. This helps robots to perform tasks with complex end-effector constraints such as grinding and welding more efficiently and autonomously.\n",
      "url": "https://www.semanticscholar.org/paper/332f84439359c8b9b288e2a00ffbcc6721a94a8b",
      "year": 2024,
      "venue": "Industrial robot",
      "source": "semantic_scholar",
      "doi": "10.1108/ir-10-2023-0251",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.011508566397619793,
      "novelty_score": 0.9562499999999999,
      "recency_score": 0.8,
      "relevance_score": 0.16385256991928598,
      "bm25_score": 0.0,
      "combined_score": 0.16385256991928598,
      "rank": 200
    },
    {
      "title": "Manifold Learning for fMRI time-varying FC",
      "authors": [
        "J. Gonzalez-Castillo",
        "Isabel S. Fernandez",
        "K. Lam",
        "D. Handwerker",
        "Francisco Pereira",
        "P. Bandettini"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/ef84c2d8052b5af43ae95579f7943c5594cf41ed",
      "year": 2023,
      "venue": "bioRxiv",
      "source": "semantic_scholar",
      "doi": "10.1101/2023.01.14.523992",
      "pdf_url": "https://www.biorxiv.org/content/biorxiv/early/2023/01/16/2023.01.14.523992.full.pdf",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.06145712627403771,
      "novelty_score": 0.9152542372881356,
      "recency_score": 0.7,
      "relevance_score": 0.15883713788221132,
      "bm25_score": 0.0,
      "combined_score": 0.15883713788221132,
      "rank": 201
    },
    {
      "title": "Integral formulas for a foliated sub-Riemannian manifold",
      "authors": [
        "V. Rovenski"
      ],
      "abstract": "We apply the notion of foliation to a nonholonomic manifold, which was introduced for the geometric interpretation of constrained systems in mechanics. We prove a series of integral formulas for a foliated sub-Riemannian manifold, that is, a Riemannian manifold equipped with a distribution $${{\\mathscr {D}}}$$ D and a foliation $${{\\mathscr {F}}}$$ F whose tangent bundle is a subbundle of  $${{\\mathscr {D}}}$$ D . Our integral formulas generalize some results for a foliated Riemannian manifold and involve the shape operators of $${\\mathscr {F}}$$ F with respect to normals in $${\\mathscr {D}}$$ D , the curvature tensor of induced connection on $${\\mathscr {D}}$$ D and arbitrary functions depending on elementary symmetric functions of eigenvalues of the shape operators. For a special choice of these functions, integral formulas with the Newton transformations of the shape operators of $${\\mathscr {F}}$$ F are obtained. Application to a foliated sub-Riemannian manifold with restrictions on the curvature and extrinsic geometry of  $${\\mathscr {F}}$$ F and also to codimension-one foliations are given.",
      "url": "https://www.semanticscholar.org/paper/cb380c2885dfb70009c60e498cd130d99c1b8595",
      "year": 2021,
      "venue": "European Journal of Mathematics",
      "source": "semantic_scholar",
      "doi": "10.1007/s40879-023-00593-5",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.055519204273252555,
      "novelty_score": 0.9122657580919931,
      "recency_score": 0.5,
      "relevance_score": 0.11705576128197577,
      "bm25_score": 0.0,
      "combined_score": 0.11705576128197577,
      "rank": 202
    },
    {
      "title": "Hyper-Laplacian Regularized Low-Rank Collaborative Representation Classification",
      "authors": [
        "Shun Xu",
        "Wenwen Shen"
      ],
      "abstract": "Face recognition is an important branch of computer vision. Domestic and foreign scholars have proposed many algorithms to improve the face recognition rate. However, when the training sample and the test sample are exposed to light, occlusion or contamination, the performance of the proposed algorithm will decrease. The recently proposed low-rank constrained collaborative representation classification algorithm (LCRC) has been proven to have superior performance in face recognition. The model is a global clustering method that can effectively recover the global subspace structures of the data, but does not consider the local geometric manifold structures of the original data. This will cause it to break the manifold structures of the original data while restoring the data, thereby losing the local geometric information of the recovered data. For the flaws of the algorithm, this paper proposes a hyper-Laplacian regularized low-rank collaborative representation classification (HLCRC). The hyper-Laplacian regularizer is introduced into the low-rank collaborative representation model to maintain the multivariate geometric manifold structures between data. Experiments on public face database show that the proposed algorithm is superior to many existing algorithms in face recognition rate.",
      "url": "https://www.semanticscholar.org/paper/d00180bafb3fffc3ee95496a0d875909379aaf3f",
      "year": 2020,
      "venue": "International Conference on Advanced Computational Intelligence",
      "source": "semantic_scholar",
      "doi": "10.1109/ICACI49185.2020.9177524",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0975778564504455,
      "novelty_score": 0.9818181818181817,
      "recency_score": 0.4,
      "relevance_score": 0.10967335693513366,
      "bm25_score": 0.0,
      "combined_score": 0.10967335693513366,
      "rank": 203
    },
    {
      "title": "mHC: Manifold-Constrained Hyper-Connections",
      "authors": [
        "Zhenda Xie",
        "Yixuan Wei",
        "Huanqi Cao",
        "Chenggang Zhao",
        "Chengqi Deng",
        "Jiashi Li",
        "Damai Dai",
        "Huazuo Gao",
        "Jiang Chang",
        "Liang Zhao"
      ],
      "abstract": "Recently, studies exemplified by Hyper-Connections (HC) have extended the ubiquitous residual connection paradigm established over the past decade by expanding the residual stream width and diversifying connectivity patterns. While yielding substantial performance gains, this diversification fundamentally compromises the identity mapping property intrinsic to the residual connection, which causes severe training instability and restricted scalability, and additionally incurs notable memory access overhead. To address these challenges, we propose Manifold-Constrained Hyper-Connections (mHC), a general framework that projects the residual connection space of HC onto a specific manifold to restore the identity mapping property, while incorporating rigorous infrastructure optimization to ensure efficiency. Empirical experiments demonstrate that mHC is effective for training at scale, offering tangible performance improvements and superior scalability. We anticipate that mHC, as a flexible and practical extension of HC, will contribute to a deeper understanding of topological architecture design and suggest promising directions for the evolution of foundational models.",
      "url": "http://arxiv.org/abs/2512.24880v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24880v1",
      "citations": null,
      "categories": [
        "cs.CL",
        "cs.AI",
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.23083566946856618,
      "novelty_score": 0.9745222929936307,
      "recency_score": 0.9,
      "relevance_score": 0.24925070084056988,
      "bm25_score": 0.0,
      "combined_score": 0.24925070084056988,
      "rank": 204
    },
    {
      "title": "Hyper-Generalized Weakly Symmetric Para-Sasakian Manifolds and Their Geometric Properties",
      "authors": [
        "B. Thangjam",
        "M. Devi"
      ],
      "abstract": "This paper examines para-Sasakian manifolds that satisfy a hyper-generalized weakly symmetric curvature condition. The conditions under which such a manifold with a hyper-generalized weakly symmetric curvature condition satisfies the η-Einstein manifold are established. Furthermore, the geometric behavior of a hyper-generalized weakly symmetric para-Sasakian manifold admitting quarter-symmetric metric connection is analyzed.",
      "url": "https://www.semanticscholar.org/paper/c673f9021fdb127edcebdf60239985331b4b693c",
      "year": 2025,
      "venue": "BULLETIN OF THE KARAGANDA UNIVERSITY-MATHEMATICS",
      "source": "semantic_scholar",
      "doi": "10.31489/2025m2/241-251",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.15487017662022132,
      "novelty_score": 0.9509668508287293,
      "recency_score": 0.9,
      "relevance_score": 0.2264610529860664,
      "bm25_score": 0.0,
      "combined_score": 0.2264610529860664,
      "rank": 205
    },
    {
      "title": "Generative Bayesian Hyperparameter Tuning",
      "authors": [
        "Hedibert Lopes",
        "Nick Polson",
        "Vadim Sokolov"
      ],
      "abstract": "\\noindent Hyper-parameter selection is a central practical problem in modern machine learning, governing regularization strength, model capacity, and robustness choices. Cross-validation is often computationally prohibitive at scale, while fully Bayesian hyper-parameter learning can be difficult due to the cost of posterior sampling. We develop a generative perspective on hyper-parameter tuning that combines two ideas: (i) optimization-based approximations to Bayesian posteriors via randomized, weighted objectives (weighted Bayesian bootstrap), and (ii) amortization of repeated optimization across many hyper-parameter settings by learning a transport map from hyper-parameters (including random weights) to the corresponding optimizer. This yields a ``generator look-up table'' for estimators, enabling rapid evaluation over grids or continuous ranges of hyper-parameters and supporting both predictive tuning objectives and approximate Bayesian uncertainty quantification. We connect this viewpoint to weighted $M$-estimation, envelope/auxiliary-variable representations that reduce non-quadratic losses to weighted least squares, and recent generative samplers for weighted $M$-estimators.",
      "url": "http://arxiv.org/abs/2512.20051v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.20051v1",
      "citations": null,
      "categories": [
        "stat.ML",
        "stat.CO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.14049412891544663,
      "novelty_score": 0.9902912621359223,
      "recency_score": 0.9,
      "relevance_score": 0.222148238674634,
      "bm25_score": 0.0,
      "combined_score": 0.222148238674634,
      "rank": 206
    },
    {
      "title": "Randers metrics with compatible linear connections: a coordinate-free approach",
      "authors": [
        "M'ark Ol'ah",
        "Csaba Vincze"
      ],
      "abstract": "A Randers space is a differentiable manifold equipped with a Randers metric. It is the sum of a Riemannian metric and a one-form on the base manifold. The compatibility of a linear connection with the metric means that the parallel transports preserve the Randers norm of tangent vectors. The existence of such a linear connection is not guaranteed in general. If it does exist then we speak about a generalized Berwald Randers metric. In what follows we give a necessary and sufficient condition for a Randers metric to be a generalized Berwald metric and we describe some distinguished compatible linear connections. The method is based on the solution of constrained optimization problems for tensors that are in one-to-one correspondence to the compatible linear connections. The solutions are given in terms of explicit formulas by choosing the free tensor components to be zero. Throughout the paper we use a coordinate-free approach to keep the geometric feature of the argumentation as far as possible.",
      "url": "https://www.semanticscholar.org/paper/ec2c66f5976abfcd2f630903c6e1a735eefa2a97",
      "year": 2025,
      "venue": "Journal of Geometry",
      "source": "semantic_scholar",
      "doi": "10.1007/s00022-025-00755-8",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.12682950962863299,
      "novelty_score": 0.9542619542619543,
      "recency_score": 0.9,
      "relevance_score": 0.2180488528885899,
      "bm25_score": 0.0,
      "combined_score": 0.2180488528885899,
      "rank": 207
    },
    {
      "title": "Arithmetic monodromy of hyper-K\\\"ahler varieties over $p$-adic fields",
      "authors": [
        "Kazuhiro Ito",
        "Tetsushi Ito",
        "Teruhisa Koshikawa",
        "Teppei Takamatsu",
        "Haitao Zou"
      ],
      "abstract": "In this paper, we study the $p$-adic and $\\ell$-adic monodromy operators associated with hyper-K\\\"ahler varieties over $p$-adic fields, in connection with Looijenga-Lunts-Verbitsky Lie algebras. We investigate a conjectural relation between the nilpotency indices of these monodromy operators on higher-degree cohomology groups and on the second cohomology, which may be viewed as an arithmetic analogue of Nagai's conjecture for degenerations of hyper-K\\\"ahler manifolds over a disk. We verify this arithmetic version of Nagai's conjecture for hyper-K\\\"ahler varieties over $p$-adic fields, assuming they belong to one of the four known deformation types. As part of our approach, we introduce a new method to analyze the $p$-adic cohomology of hyper-K\\\"ahler varieties via Sen's theory.",
      "url": "https://www.semanticscholar.org/paper/73641545c3fda2c186b4b41aa3e27f19d7d96bc6",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.1104870172450309,
      "novelty_score": 0.955503512880562,
      "recency_score": 0.9,
      "relevance_score": 0.2131461051735093,
      "bm25_score": 0.0,
      "combined_score": 0.2131461051735093,
      "rank": 208
    },
    {
      "title": "A Single-Loop First-Order Algorithm for Linearly Constrained Bilevel Optimization",
      "authors": [
        "Wei Shen",
        "Jiawei Zhang",
        "Minhui Huang",
        "Cong Shen"
      ],
      "abstract": "We study bilevel optimization problems where the lower-level problems are strongly convex and have coupled linear constraints. To overcome the potential non-smoothness of the hyper-objective and the computational challenges associated with the Hessian matrix, we utilize penalty and augmented Lagrangian methods to reformulate the original problem as a single-level one. Especially, we establish a strong theoretical connection between the reformulated function and the original hyper-objective by characterizing the closeness of their values and derivatives. Based on this reformulation, we propose a single-loop, first-order algorithm for linearly constrained bilevel optimization (SFLCB). We provide rigorous analyses of its non-asymptotic convergence rates, showing an improvement over prior double-loop algorithms -- form $O(\\epsilon^{-3}\\log(\\epsilon^{-1}))$ to $O(\\epsilon^{-3})$. The experiments corroborate our theoretical findings and demonstrate the practical efficiency of the proposed SFLCB algorithm. Simulation code is provided at https://github.com/ShenGroup/SFLCB.",
      "url": "https://www.semanticscholar.org/paper/9db6d6af5761c0e98bac91d2cf09af65880ca0e9",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2510.24710",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.07529097776818147,
      "novelty_score": 0.9247817327065144,
      "recency_score": 0.9,
      "relevance_score": 0.20258729333045447,
      "bm25_score": 0.0,
      "combined_score": 0.20258729333045447,
      "rank": 209
    },
    {
      "title": "Lagrangian Dual Sections: A Topological Perspective on Hidden Convexity",
      "authors": [
        "Venkat Chandrasekaran",
        "Timothy Duff",
        "Jose Israel Rodriguez",
        "Kevin Shu"
      ],
      "abstract": "Hidden convexity is a powerful idea in optimization: under the right transformations, nonconvex problems that are seemingly intractable can be solved efficiently using convex optimization. We introduce the notion of a Lagrangian dual section of a nonlinear program defined over a topological space, and we use it to give a sufficient condition for a nonconvex optimization problem to have a natural convex reformulation. We emphasize the topological nature of our framework, using only continuity and connectedness properties of a certain Lagrangian formulation of the problem to prove our results. We demonstrate the practical consequences of our framework in a range of applications and by developing new algorithmic methodology. First, we present families of nonconvex problem instances that can be transformed to convex programs in the context of spectral inverse problems -- which include quadratically constrained quadratic optimization and Stiefel manifold optimization as special cases -- as well as unbalanced Procrustes problems. In each of these applications, we both generalize prior results on hidden convexity and provide unifying proofs. For the case of the spectral inverse problems, we also present a Lie-theoretic approach that illustrates connections with the Kostant convexity theorem. Second, we introduce new algorithmic ideas that can be used to find globally optimal solutions to both Lagrangian forms of an optimization problem as well as constrained optimization problems when the underlying topological space is a Riemannian manifold.",
      "url": "https://www.semanticscholar.org/paper/702a8cf538d6f51d0b98a89a34a5d1d55b40b3ac",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.06588701814082804,
      "novelty_score": 0.9529411764705883,
      "recency_score": 0.9,
      "relevance_score": 0.19976610544224843,
      "bm25_score": 0.0,
      "combined_score": 0.19976610544224843,
      "rank": 210
    },
    {
      "title": "Hyperk\\\"ahler structures on leaves of hyper-Lie Poisson manifolds",
      "authors": [
        "Dadi Ni",
        "Kaichuan Qi"
      ],
      "abstract": "Due to its rich structure and close connection with gauge theory, hyperk\\\"ahler manifolds have attracted increasing interest. Using infinite dimensional hyperk\\\"ahler reduction, Kronheimer proved that certain adjoint orbits of complexified semisimple Lie algebras admits hyperk\\\"ahler structures. Later on, Xu obtained a proof for the existence of hyperk\\\"ahler structures on adjoint orbits of $\\mathfrak{sl}(2,\\mathbb{C})$ from the viewpoint of symplectic geometry. This paper aims to thoroughly investigate and elucidate the key differences as well as the underlying connections between two distinct construction methods.",
      "url": "https://www.semanticscholar.org/paper/786f152b5c97bc7c34f3dd3bad4a694af13c23d8",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.06533738531511074,
      "novelty_score": 0.9329268292682927,
      "recency_score": 0.9,
      "relevance_score": 0.19960121559453325,
      "bm25_score": 0.0,
      "combined_score": 0.19960121559453325,
      "rank": 211
    },
    {
      "title": "On the rigidity of special and exceptional geometries with torsion a closed $3$-form",
      "authors": [
        "Georgios Papadopoulos"
      ],
      "abstract": "We demonstrate that all Riemannian manifolds $(M, g, H)$ that admit a connection $\\hat\\nabla$ with torsion a 3-form $H$, which is both closed $d H=0$ and $\\hat\\nabla$-covariantly constant, are locally isometric to a product $N\\times G$, where $G$ is a semisimple group and $N$ is a Riemannian manifold with $ι_V H=0$ for all tangent vectors $V \\in T_pN\\subset T_pM$, $p\\in M$. If $M$ is simply connected and complete, then by the de Rham theorem $M=N\\times G$ globally. We use this to simplify the proof of similar results for strong KT, CYT and HKT manifolds that obey the above hypotheses and extend them to strong $G_2$ and $\\mathrm{Spin}(7)$ manifolds with torsion. As an application, we describe the geometry of all complete and simply connected $G_2$ and $\\mathrm{Spin}(7)$ manifolds whose torsion satisfies the above conditions.\n  We also demonstrate that all compact 8-dimensional manifolds with strong HKT structure are locally isometric to one of the following: 8-dimensional hyper-Kähler; $SU(3)$ equipped with the bi-invariant metric and 3-form; or the product $(U(1)\\times SU(2))\\times B^4$, where $B^4$ is either a hyper-Kähler manifold or $U(1)\\times SU(2)$ equipped with an HKT structure.",
      "url": "http://arxiv.org/abs/2511.20568v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.20568v2",
      "citations": null,
      "categories": [
        "math.DG",
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.06313558716507131,
      "novelty_score": 0.9012233801540552,
      "recency_score": 0.9,
      "relevance_score": 0.1989406761495214,
      "bm25_score": 0.0,
      "combined_score": 0.1989406761495214,
      "rank": 212
    },
    {
      "title": "Fundamental Limitations of QAOA on Constrained Problems and a Route to Exponential Enhancement",
      "authors": [
        "Chinonso Onah",
        "Kristel Michielsen"
      ],
      "abstract": "We study fundamental limitations of the generic Quantum Approximate Optimization Algorithm (QAOA) on constrained problems where valid solutions form a low dimensional manifold inside the Boolean hypercube, and we present a provable route to exponential improvements via constraint embedding. Focusing on permutation constrained objectives, we show that the standard generic QAOA ansatz, with a transverse field mixer and diagonal r local cost, faces an intrinsic feasibility bottleneck: even after angle optimization, circuits whose depth grows at most linearly with n cannot raise the total probability mass on the feasible manifold much above the uniform baseline suppressed by the size of the full Hilber space. Against this envelope we introduce a minimal constraint enhanced kernel (CE QAOA) that operates directly inside a product one hot subspace and mixes with a block local XY Hamiltonian. For permutation constrained problems, we prove an angle robust, depth matched exponential enhancement where the ratio between the feasible mass from CE QAOA and generic QAOA grows exponentially in $n^2$ for all depths up to a linear fraction of n, under a mild polynomial growth condition on the interaction hypergraph. Thanks to the problem algorithm co design in the kernel construction, the techniques and guarantees extend beyond permutations to a broad class of NP-Hard constrained optimization problems.",
      "url": "http://arxiv.org/abs/2511.17259v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.17259v1",
      "citations": null,
      "categories": [
        "quant-ph",
        "cs.CC",
        "cs.CE",
        "cs.DM",
        "math-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.05770646306971069,
      "novelty_score": 0.9187066974595843,
      "recency_score": 0.9,
      "relevance_score": 0.19731193892091323,
      "bm25_score": 0.0,
      "combined_score": 0.19731193892091323,
      "rank": 213
    },
    {
      "title": "Lie algebroids, quantum Poisson algebroids, and Lie algebroid connections",
      "authors": [
        "Satyendra Kumar Mishra",
        "A. Sarkar"
      ],
      "abstract": "In this paper, we consider Lie algebroids over commutative ringed spaces. Lie algebroids over ringed spaces unify the existing notion of Lie algebroids over smooth manifolds, complex manifolds, analytic spaces, algebraic varieties, and schemes. We show that the universal enveloping algebroid of a Lie algebroid possesses a natural filtration that yields a structure of a sheaf of quantum Poisson algebras. We establish a bijective correspondence between sheaves of quantum Poisson algebras and Lie algebroids. We show that this correspondence leads to an adjunction between the two categories. We discuss this bijective correspondence in particular cases of Lie algebroids over ringed spaces and highlight the subsequent results. To characterize non-flat Lie algebroid connections, we construct a sheaf of twisted universal enveloping algebras for a Lie algebroid using Lie algebroid (hyper) cohomology. We show that our construction yields some of the existing constructions for Lie-Rinehart algebras and holomorphic Lie algebroids. As another application, we study the deformation groupoid of a Lie algebroid using the second hypercohomology of the Lie algebroid.",
      "url": "https://www.semanticscholar.org/paper/4105091889345f991b2875d21437a69e73fd6b4e",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.05463036426078386,
      "novelty_score": 0.9470426409903714,
      "recency_score": 0.9,
      "relevance_score": 0.19638910927823516,
      "bm25_score": 0.0,
      "combined_score": 0.19638910927823516,
      "rank": 214
    },
    {
      "title": "Trajectory Optimization by Successive Pseudospectral Convexification on Riemannian Manifolds",
      "authors": [
        "Tatsuya Narumi",
        "Shin-ichiro Sakai"
      ],
      "abstract": "This paper proposes an intrinsic pseudospectral convexification framework for optimal control problems with manifold constraints. While successive pseudospectral convexification combines spectral collocation with successive convexification, classical pseudospectral methods are not geometry-consistent on manifolds. This is because interpolation and differentiation are performed in Euclidean coordinates. We introduce a geometry-consistent transcription that enables pseudospectral collocation without imposing manifold constraints extrinsically. The resulting method solves nonconvex manifold-constrained problems through a sequence of convex subproblems. A six-degree-of-freedom landing guidance example with unit quaternions and unit thrust-direction vectors demonstrates the practicality of the approach and preserves manifold feasibility to machine precision.",
      "url": "http://arxiv.org/abs/2512.09551v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.09551v1",
      "citations": null,
      "categories": [
        "math.OC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.05427616791745606,
      "novelty_score": 0.9609211444521981,
      "recency_score": 0.9,
      "relevance_score": 0.19628285037523685,
      "bm25_score": 0.0,
      "combined_score": 0.19628285037523685,
      "rank": 215
    },
    {
      "title": "Cech - de Rham Chern character on the stack of holomorphic vector bundles",
      "authors": [
        "Cheyne Glass",
        "T. Tradler",
        "M. Zeinalian"
      ],
      "abstract": "We provide a formula for the Chern character of a holomorphic vector bundle in the hyper-cohomology of the de Rham complex of holomorphic sheaves on a complex manifold. This Chern character can be thought of as a completion of the Chern character in Hodge cohomology obtained as the trace of the exponential of the Atiyah class, which is \\v{C}ech closed, to one that is \\v{C}ech-Del closed. Such a completion is a key step toward lifting O'Brian-Toledo-Tong invariants of coherent sheaves from Hodge cohomology to de Rham cohomology. An alternate approach toward the same end goal, instead using simplicial differential forms and Green complexes, can be found in Hosgood's works [Ho1, Ho2]. In the algebraic setting, and more generally for K\\\"{a}hler manifolds, where Hodge and de Rham cohomologies agree, such extensions are not necessary, whereas in the non-K\\\"{a}hler, or equivariant settings the two theories differ. We provide our formulae as a map of simplicial presheaves, which readily extend the results to the equivariant setting and beyond. This paper can be viewed as a sequel to [GMTZ1] which covered such a discussion in Hodge cohomology. As an aside, we give a conceptual understanding of how formulas obtained by Bott and Tu for Chern classes using transition functions and those from Chern-Weil theory using connections, are part of a natural unifying story.",
      "url": "https://www.semanticscholar.org/paper/1231982e3a59d14420806f9e0aa5b6d34d2537d5",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.051699793955580844,
      "novelty_score": 0.9422074846044529,
      "recency_score": 0.9,
      "relevance_score": 0.19550993818667428,
      "bm25_score": 0.0,
      "combined_score": 0.19550993818667428,
      "rank": 216
    },
    {
      "title": "The algebraic square of an irreducible complex spinor",
      "authors": [
        "Alejandro Gil-Garc'ia",
        "C. Shahbazi"
      ],
      "abstract": "We characterize, in every dimension and signature, the algebraic squares of an irreducible complex spinor as a pair of exterior forms satisfying a prescribed system of algebraic relations that we present in terms of the geometric product of the underlying quadratic vector space. As a result, we obtain a general correspondence between irreducible complex spinors and algebraically constrained exterior forms, which clarifies the subtle relationship between spinors and exterior forms and contributes towards the understanding of spinors as the square root of geometry. We use this formalism to construct the squares of an irreducible complex spinor in Euclidean dimensions up to six, and also to construct the squares of a generic, possibly non-pure and non-unit, irreducible complex chiral spinor in eight Euclidean dimensions. Elaborating on this result, we consider a natural notion of spinorial instanton that we study for connections on a principal bundle with a complex structure group as well as for curvings of a $\\mathbb{C}^{\\ast}$-bundle gerbe defined on a Lorentzian six-manifold.",
      "url": "https://www.semanticscholar.org/paper/7a97a602f138d4259eac9619fc1948877dca3b10",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.04789229056371509,
      "novelty_score": 0.9300911854103344,
      "recency_score": 0.9,
      "relevance_score": 0.19436768716911454,
      "bm25_score": 0.0,
      "combined_score": 0.19436768716911454,
      "rank": 217
    },
    {
      "title": "Quotient Manifold Optimization for Spectral Compressed Sensing",
      "authors": [
        "Wenlong Wang",
        "Wen Huang",
        "Zai Yang"
      ],
      "abstract": "Spectral compressed sensing involves reconstructing a spectral-sparse signal from a subset of uniformly spaced samples, with applications in radar imaging and wireless channel estimation. By fully exploiting the signal structures, this problem is formulated as a rank-constrained semidefinite program subject to Hankel-Toeplitz structural constraints in our previous work. To further enhance computational efficiency, this paper proposes a quotient-manifold-based optimization framework that leverages the underlying Riemannian geometry in a matrix factorization space. Specifically, we establish an equivalence between spectral-sparse signals and matrix equivalence classes under the action of the real orthogonal group, where each class member corresponds to a rank-constrained positive-semidefinite Hankel-Toeplitz structured matrix. The associated quotient manifold geometry--including the Riemannian metric, horizontal space, retraction, and vector transport--is rigorously derived. Based on these results, we develop a Riemannian conjugate gradient descent algorithm, where each iteration is efficiently implemented using fast Fourier transforms (FFTs) by exploiting the Hankel and Toeplitz structures. Extensive numerical experiments demonstrate the superior performance of the proposed algorithm in both computational speed and accuracy compared to state-of-the-art methods.",
      "url": "http://arxiv.org/abs/2511.19108v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.19108v1",
      "citations": null,
      "categories": [
        "math.OC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.046892177404137925,
      "novelty_score": 0.931304347826087,
      "recency_score": 0.9,
      "relevance_score": 0.1940676532212414,
      "bm25_score": 0.0,
      "combined_score": 0.1940676532212414,
      "rank": 218
    },
    {
      "title": "Signatures of real-space geometry, topology, and metric tensor in quantum transport in periodically corrugated spaces",
      "authors": [
        "Benjamin Schwager",
        "Theresa Appel",
        "Jamal Berakdar"
      ],
      "abstract": "The motion of a quantum particle constrained to a two-dimensional non-compact Riemannian manifold with non-trivial metric can be described by a flat-space Schroedinger-type equation at the cost of introducing local mass and metric and geometry-induced effective potential with no classical counterpart. For a metric tensor periodically modulated along one dimension, the formation of bands is demonstrated and transport-related quantities are derived. Using S-matrix approach, the quantum conductance along the manifold is calculated and contrasted with conventional quantum transport methods in flat spaces. The topology, e.g. whether the manifold is simply connected, compact or non-compact shows up in global, non-local properties such as the Aharonov-Bohm phase. The results vividly demonstrate emergent phenomena due to the interplay of reduced-dimensionality, particles quantum nature, geometry, and topology.",
      "url": "http://arxiv.org/abs/2512.16846v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.16846v1",
      "citations": null,
      "categories": [
        "cond-mat.mes-hall",
        "math-ph",
        "physics.class-ph",
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.04553233842426067,
      "novelty_score": 0.9291497975708501,
      "recency_score": 0.9,
      "relevance_score": 0.1936597015272782,
      "bm25_score": 0.0,
      "combined_score": 0.1936597015272782,
      "rank": 219
    },
    {
      "title": "Secrets of the Goo: The genome assembly of the Pacific banana slug, Ariolimax columbianus",
      "authors": [
        "Max Genetti",
        "Merly Escalona",
        "Cade Mirchandani",
        "Jonas Oppenheimer",
        "E. Beraut",
        "Samuel Sacco",
        "William E. Seligmann",
        "Colin Fairbairn",
        "Ruta Sahasrabudhe",
        "Mohan P A Marimuthu"
      ],
      "abstract": "Abstract The Pacific banana slug, Ariolimax columbianus, is endemic to the forests of the Pacific Northern West. Found throughout the coastal foothills and mountains of California, the hermaphroditic molluscs Ariolimax spp. are niche-constrained, hyper-localized, and phenotypically diverse. The evolutionary history, recent population history and environmental conditions leading to their phenotypic and genetic variation are not understood. To facilitate such research, we present the first high-quality de novo genome assembly of A. columbianus as part of the California Conservation Genomics Project. Pacific Biosciences HiFi long reads and Omni-C chromatin-proximity sequencing technologies were used to produce a de novo genome assembly, consistent with the standard California Conservation Genomics Project genome assembly protocol. This assembly comprises 401 scaffolds spanning 2.29 Gb, represented by a scaffold N50 of 94.9 Mb, a contig N50 of 3.7 Mb, and a benchmarking universal single-copy ortholog completeness score of 93.9%. Future work will use the A. columbianus genome to study the population structure of Ariolimax spp. across California to understand patterns of population structure, genetic diversity, and the broader ecological connections with their habitat. This data will contribute to the California Conservation Genomics Project, expanding the knowledge about the partitioning of genomic variation across the different ecoregions of California.",
      "url": "https://www.semanticscholar.org/paper/29dea4859e36ffe56831ffd1850a48666ae0b38e",
      "year": 2025,
      "venue": "Journal of Heredity",
      "source": "semantic_scholar",
      "doi": "10.1093/jhered/esaf002",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.04452309643757406,
      "novelty_score": 0.9107142857142856,
      "recency_score": 0.9,
      "relevance_score": 0.19335692893127224,
      "bm25_score": 0.0,
      "combined_score": 0.19335692893127224,
      "rank": 220
    },
    {
      "title": "Physics-Constrained Neural Dynamics: A Unified Manifold Framework for Large-Scale Power Flow Computation",
      "authors": [
        "Xuezhi Liu"
      ],
      "abstract": "Power flow analysis is a fundamental tool for power system analysis, planning, and operational control. Traditional Newton-Raphson methods suffer from limitations such as initial value sensitivity and low efficiency in batch computation, while existing deep learning-based power flow solvers mostly rely on supervised learning, requiring pre-solving of numerous cases and struggling to guarantee physical consistency. This paper proposes a neural physics power flow solving method based on manifold geometry and gradient flow, by describing the power flow equations as a constraint manifold, and constructing an energy function \\(V(\\mathbf{x}) = \\frac{1}{2}\\|\\mathbf{F}(\\mathbf{x})\\|^2\\) and gradient flow \\(\\frac{d\\mathbf{x}}{dt} = -\\nabla V(\\mathbf{x})\\), transforming power flow solving into an equilibrium point finding problem for dynamical systems. Neural networks are trained in an unsupervised manner by directly minimizing physical residuals, requiring no labeled data, achieving true \"end-to-end\" physics-constrained learning.",
      "url": "http://arxiv.org/abs/2512.01207v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.01207v1",
      "citations": null,
      "categories": [
        "eess.SY",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.044404756049423934,
      "novelty_score": 0.9315068493150686,
      "recency_score": 0.9,
      "relevance_score": 0.1933214268148272,
      "bm25_score": 0.0,
      "combined_score": 0.1933214268148272,
      "rank": 221
    },
    {
      "title": "Semantic Geometry for policy-constrained interpretation",
      "authors": [
        "Nikit Phadke"
      ],
      "abstract": "We present a geometric framework for policy-constrained semantic interpretation that provably prevents hallucinated commitments in high-stakes domains. Semantic meaning is represented as direction on a unit sphere, evidence is modeled as sets of witness vectors, and admissible interpretations correspond to spherical convex regions. Policy constraints are introduced as explicit priors defined over the same manifold, separated from evidence geometry. Interpretation reduces to constrained optimization over admissible regions, with refusal emerging as a topologically necessary outcome under contradiction or policy exclusion. We connect this framework to information theory, Bayesian inference, and sheaf-theoretic semantics, proving that our complexity bounds are information-theoretically optimal. Empirical validation on large scale regulated financial data demonstrates zero hallucinated approvals across multiple policy regimes-the first such result at scale.",
      "url": "http://arxiv.org/abs/2512.14731v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.14731v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.04324404132327859,
      "novelty_score": 0.92057761732852,
      "recency_score": 0.9,
      "relevance_score": 0.1929732123969836,
      "bm25_score": 0.0,
      "combined_score": 0.1929732123969836,
      "rank": 222
    },
    {
      "title": "Quantumness via Discrete Structures",
      "authors": [
        "Ravi Kunjwal"
      ],
      "abstract": "Quantum theory departs from classical probabilistic theories in foundational ways. These departures--termed quantumness here--power quantum information and computation. This thesis charts the role of discrete structures in assessing quantumness, synthesizing elements of my postdoctoral research through this lens. After an introduction to the necessary background concepts, I present my work under three broad categories. First, I present work on contextuality that extensively relies on (undirected) graphs and hypergraphs as the discrete structures of interest; more specifically, it relies on invariants associated with them. This work includes Kochen-Specker (KS) contextuality and its operationalization to generalized contextuality, expressed via (hyper)graph-theoretic frameworks. I also present work on KS-contextuality in multiqubit systems and an application of generalized contextuality to a one-shot communication task, both of which rely on hypergraphs. Second, I present work on causality, where the discrete structures of interest are directed graphs. This includes work on indefinite causal order, specifically its connections to the gap between local operations and classical communication (LOCC) and separable operations (SEP), and a device-independent notion of nonclassicality--termed antinomicity--that generalizes Bell nonlocality without global causal assumptions. Finally, I present work on the incompatibility of quantum measurements, its connection to Bell nonlocality, and its role in discriminating between quantum and almost quantum correlations in the single-system setting. The discrete structures of interest here are hypergraphs that model joint measurability relations between quantum measurements. I conclude with a summary and an overview of work that is not covered in this thesis.",
      "url": "http://arxiv.org/abs/2512.10063v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.10063v2",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0430037093403817,
      "novelty_score": 0.9823434991974318,
      "recency_score": 0.9,
      "relevance_score": 0.19290111280211453,
      "bm25_score": 0.0,
      "combined_score": 0.19290111280211453,
      "rank": 223
    },
    {
      "title": "Gradient-descent methods for quantum detector tomography",
      "authors": [
        "Amanuel Anteneh",
        "Olivier Pfister"
      ],
      "abstract": "We present a technique for performing quantum detector tomography (QDT) of phase insensitive quantum detectors using gradient descent-based optimization to learn the positive operator-valued measure (POVM) that best describes the data collected using the detector under study. We numerically benchmark our method against constrained convex optimization (CCO) and show that it reaches higher or comparable reconstruction fidelity in much less time even in the presence of noise and limited probe state resources. We also present a possible extension of our approach to the phase sensitive case via a parametrization of POVMs on the complex Stiefel manifold which enables gradient based optimization restricted to this manifold.",
      "url": "http://arxiv.org/abs/2511.14579v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.14579v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.04284093238957086,
      "novelty_score": 0.9254032258064515,
      "recency_score": 0.9,
      "relevance_score": 0.19285227971687127,
      "bm25_score": 0.0,
      "combined_score": 0.19285227971687127,
      "rank": 224
    },
    {
      "title": "AL-Net: Adaptive Learning for Enhanced Cell Nucleus Segmentation in Pathological Images",
      "authors": [
        "Zhuping Chen",
        "Sheng-Lung Peng",
        "Rui Yang",
        "Ming Zhao",
        "Chaolin Zhang"
      ],
      "abstract": "Precise segmentation of cell nuclei in pathological images is the foundation of cancer diagnosis and quantitative analysis, but blurred boundaries, scale variability, and staining differences have long constrained its reliability. To address this, this paper proposes AL-Net—an adaptive learning network that breaks through these bottlenecks through three innovative mechanisms: First, it integrates dilated convolutions with attention-guided skip connections to dynamically integrate multi-scale contextual information, adapting to variations in cell nucleus morphology and size. Second, it employs self-scheduling loss optimization: during the initial training phase, it focuses on region segmentation (Dice loss) and later switches to a boundary refinement stage, introducing gradient manifold constraints to sharpen edge localization. Finally, it designs an adaptive optimizer strategy, leveraging symbolic exploration (Lion) to accelerate convergence, and switches to gradient fine-tuning after reaching a dynamic threshold to stabilize parameters. On the 2018 Data Science Bowl dataset, AL-Net achieved state-of-the-art performance (Dice coefficient 92.96%, IoU 86.86%), reducing boundary error by 15% compared to U-Net/DeepLab; in cross-domain testing (ETIS/ColonDB polyp segmentation), it demonstrated over 80% improvement in generalization performance. AL-Net establishes a new adaptive learning paradigm for computational pathology, significantly enhancing diagnostic reliability.",
      "url": "https://www.semanticscholar.org/paper/85fa9539b6139e5c61a9364cb288675c0ae7e542",
      "year": 2025,
      "venue": "Electronics",
      "source": "semantic_scholar",
      "doi": "10.3390/electronics14173507",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.04198349194052661,
      "novelty_score": 0.9407490217998882,
      "recency_score": 0.9,
      "relevance_score": 0.192595047582158,
      "bm25_score": 0.0,
      "combined_score": 0.192595047582158,
      "rank": 225
    },
    {
      "title": "Necessary and sufficient conditions for high dimensional Central Limit Theorem under moment conditions",
      "authors": [
        "Debraj Das",
        "Soumendra Lahiri"
      ],
      "abstract": "High dimensional central limit theorems (the CLTs) have been extensively studied in recent years under a variety of sufficient moment conditions connecting the dimension growth rate with the tail decay rate. In this article, we investigate whether the existing moment conditions are also necessary under the independence of the components. We consider four exhaustive classes, viz. when underlying random variables (I) have all polynomial moments, (II) have some polynomial moment of order higher than two, (III) have only second moment but no polynomial moment higher than two exists, and (IV) have infinite second moment, but belong to the domain of attraction of normal distribution. We find the optimal growth rate of the dimension with respect to sample size in the high dimensional CLTs over hyper-rectangles. More precisely, we derive necessary and sufficient moment conditions for the validity of the the CLT over hyper-rectangles in each of the four regimes listed above, showing that the CLT may hold under much weaker conditions compared to those considered in the existing literature.",
      "url": "http://arxiv.org/abs/2512.22312v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.22312v1",
      "citations": null,
      "categories": [
        "math.PR"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.04071593606933595,
      "novelty_score": 0.9435483870967741,
      "recency_score": 0.9,
      "relevance_score": 0.19221478082080082,
      "bm25_score": 0.0,
      "combined_score": 0.19221478082080082,
      "rank": 226
    },
    {
      "title": "ManifoldFormer: Geometric Deep Learning for Neural Dynamics on Riemannian Manifolds",
      "authors": [
        "Yihang Fu",
        "Lifang He",
        "Qingyu Chen"
      ],
      "abstract": "Existing EEG foundation models mainly treat neural signals as generic time series in Euclidean space, ignoring the intrinsic geometric structure of neural dynamics that constrains brain activity to low-dimensional manifolds. This fundamental mismatch between model assumptions and neural geometry limits representation quality and cross-subject generalization. ManifoldFormer addresses this limitation through a novel geometric deep learning framework that explicitly learns neural manifold representations. The architecture integrates three key innovations: a Riemannian VAE for manifold embedding that preserves geometric structure, a geometric Transformer with geodesic-aware attention mechanisms operating directly on neural manifolds, and a dynamics predictor leveraging neural ODEs for manifold-constrained temporal evolution. Extensive evaluation across four public datasets demonstrates substantial improvements over state-of-the-art methods, with 4.6-4.8% higher accuracy and 6.2-10.2% higher Cohen's Kappa, while maintaining robust cross-subject generalization. The geometric approach reveals meaningful neural patterns consistent with neurophysiological principles, establishing geometric constraints as essential for effective EEG foundation models.",
      "url": "http://arxiv.org/abs/2511.16828v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.16828v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.040709476419229014,
      "novelty_score": 0.9156193895870737,
      "recency_score": 0.9,
      "relevance_score": 0.19221284292576873,
      "bm25_score": 0.0,
      "combined_score": 0.19221284292576873,
      "rank": 227
    },
    {
      "title": "Local Path Optimization in The Latent Space Using Learned Distance Gradient",
      "authors": [
        "Jiawei Zhang",
        "Chengchao Bai",
        "Wei Pan",
        "Tianhang Liu",
        "Jifeng Guo"
      ],
      "abstract": "Constrained motion planning is a common but challenging problem in robotic manipulation. In recent years, data-driven constrained motion planning algorithms have shown impressive planning speed and success rate. Among them, the latent motion method based on manifold approximation is the most efficient planning algorithm. Due to errors in manifold approximation and the difficulty in accurately identifying collision conflicts within the latent space, time-consuming path validity checks and path replanning are required. In this paper, we propose a method that trains a neural network to predict the minimum distance between the robot and obstacles using latent vectors as inputs. The learned distance gradient is then used to calculate the direction of movement in the latent space to move the robot away from obstacles. Based on this, a local path optimization algorithm in the latent space is proposed, and it is integrated with the path validity checking process to reduce the time of replanning. The proposed method is compared with state-of-the-art algorithms in multiple planning scenarios, demonstrating the fastest planning speed",
      "url": "http://arxiv.org/abs/2512.24272v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": "10.1109/IROS60139.2025.11247535",
      "pdf_url": "https://arxiv.org/pdf/2512.24272v1",
      "citations": null,
      "categories": [
        "cs.RO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.03667851063710253,
      "novelty_score": 0.9476351351351352,
      "recency_score": 0.9,
      "relevance_score": 0.19100355319113077,
      "bm25_score": 0.0,
      "combined_score": 0.19100355319113077,
      "rank": 228
    },
    {
      "title": "Guided Path Sampling: Steering Diffusion Models Back on Track with Principled Path Guidance",
      "authors": [
        "Haosen Li",
        "Wenshuo Chen",
        "Shaofeng Liang",
        "Lei Wang",
        "Haozhe Jia",
        "Yutao Yue"
      ],
      "abstract": "Iterative refinement methods based on a denoising-inversion cycle are powerful tools for enhancing the quality and control of diffusion models. However, their effectiveness is critically limited when combined with standard Classifier-Free Guidance (CFG). We identify a fundamental limitation: CFG's extrapolative nature systematically pushes the sampling path off the data manifold, causing the approximation error to diverge and undermining the refinement process. To address this, we propose Guided Path Sampling (GPS), a new paradigm for iterative refinement. GPS replaces unstable extrapolation with a principled, manifold-constrained interpolation, ensuring the sampling path remains on the data manifold. We theoretically prove that this correction transforms the error series from unbounded amplification to strictly bounded, guaranteeing stability. Furthermore, we devise an optimal scheduling strategy that dynamically adjusts guidance strength, aligning semantic injection with the model's natural coarse-to-fine generation process. Extensive experiments on modern backbones like SDXL and Hunyuan-DiT show that GPS outperforms existing methods in both perceptual quality and complex prompt adherence. For instance, GPS achieves a superior ImageReward of 0.79 and HPS v2 of 0.2995 on SDXL, while improving overall semantic alignment accuracy on GenEval to 57.45%. Our work establishes that path stability is a prerequisite for effective iterative refinement, and GPS provides a robust framework to achieve it.",
      "url": "http://arxiv.org/abs/2512.22881v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.22881v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.032115471396838985,
      "novelty_score": 0.9622641509433961,
      "recency_score": 0.9,
      "relevance_score": 0.18963464141905173,
      "bm25_score": 0.0,
      "combined_score": 0.18963464141905173,
      "rank": 229
    },
    {
      "title": "On Weinstein domains in symplectic manifolds",
      "authors": [
        "Thomas E. Mark",
        "Bülent Tosun"
      ],
      "abstract": "We prove that a Weinstein domain symplectically embedded in a closed symplectic manifold always admits symplectic hypersurfaces in its complement, possibly after a deformation. As a consequence, we obtain an obstruction for a closed 3-dimensional manifold to arise as the boundary of a Weinstein domain in a class of symplectic 4-manifolds that includes many symplectic rational surfaces. A particular application is that no Brieskorn homology sphere bounds a Weinstein domain symplectically embedded in a rational surface diffeomorphic to $S^2\\times S^2$ or to ${\\mathbb C} P^2\\# k \\overline{{\\mathbb C}P}^2$, for any $k\\leq 7$, despite the fact that many Brieskorn spheres bound Stein domains holomorphically embedded in these rational surfaces. Several families of Brieskorn spheres are obtained that do not bound a Weinstein domain in any 4-manifold with a ``positive'' symplectic structure. Such Weinstein domains do exist in certain positive symplectic rational surfaces when $k\\geq 8$, though their topology is significantly constrained.",
      "url": "http://arxiv.org/abs/2512.04278v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.04278v1",
      "citations": null,
      "categories": [
        "math.SG",
        "math.CV",
        "math.GT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.031825719575550304,
      "novelty_score": 0.9348268839103869,
      "recency_score": 0.9,
      "relevance_score": 0.1895477158726651,
      "bm25_score": 0.0,
      "combined_score": 0.1895477158726651,
      "rank": 230
    },
    {
      "title": "Deep Manifold Part 2: Neural Network Mathematics",
      "authors": [
        "Max Y. Ma",
        "Gen-Hua Shi"
      ],
      "abstract": "This work develops the global equations of neural networks through stacked piecewise manifolds, fixed-point theory, and boundary-conditioned iteration. Once fixed coordinates and operators are removed, a neural network appears as a learnable numerical computation shaped by manifold complexity, high-order nonlinearity, and boundary conditions. Real-world data impose strong data complexity, near-infinite scope, scale, and minibatch fragmentation, while training dynamics produce learning complexity through shifting node covers, curvature accumulation, and the rise and decay of plasticity. These forces constrain learnability and explain why capability emerges only when fixed-point regions stabilize. Neural networks do not begin with fixed points; they construct them through residual-driven iteration. This perspective clarifies the limits of monolithic models under geometric and data-induced plasticity and motivates architectures and federated systems that distribute manifold complexity across many elastic models, forming a coherent world-modeling framework grounded in geometry, algebra, fixed points, and real-data complexity.",
      "url": "http://arxiv.org/abs/2512.06563v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.06563v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.03099896404129549,
      "novelty_score": 0.9683544303797468,
      "recency_score": 0.9,
      "relevance_score": 0.18929968921238866,
      "bm25_score": 0.0,
      "combined_score": 0.18929968921238866,
      "rank": 231
    },
    {
      "title": "DAE-HardNet: A Physics Constrained Neural Network Enforcing Differential-Algebraic Hard Constraints",
      "authors": [
        "Rahul Golder",
        "Bimol Nath Roy",
        "M. M. Faruque Hasan"
      ],
      "abstract": "Traditional physics-informed neural networks (PINNs) do not always satisfy physics based constraints, especially when the constraints include differential operators. Rather, they minimize the constraint violations in a soft way. Strict satisfaction of differential-algebraic equations (DAEs) to embed domain knowledge and first-principles in data-driven models is generally challenging. This is because data-driven models consider the original functions to be black-box whose derivatives can only be obtained after evaluating the functions. We introduce DAE-HardNet, a physics-constrained (rather than simply physics-informed) neural network that learns both the functions and their derivatives simultaneously, while enforcing algebraic as well as differential constraints. This is done by projecting model predictions onto the constraint manifold using a differentiable projection layer. We apply DAE-HardNet to several systems and test problems governed by DAEs, including the dynamic Lotka-Volterra predator-prey system and transient heat conduction. We also show the ability of DAE-HardNet to estimate unknown parameters through a parameter estimation problem. Compared to multilayer perceptrons (MLPs) and PINNs, DAE-HardNet achieves orders of magnitude reduction in the physics loss while maintaining the prediction accuracy. It has the added benefits of learning the derivatives which improves the constrained learning of the backbone neural network prior to the projection layer. For specific problems, this suggests that the projection layer can be bypassed for faster inference. The current implementation and codes are available at https://github.com/SOULS-TAMU/DAE-HardNet.",
      "url": "http://arxiv.org/abs/2512.05881v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.05881v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.02829427547519147,
      "novelty_score": 0.9586466165413533,
      "recency_score": 0.9,
      "relevance_score": 0.18848828264255746,
      "bm25_score": 0.0,
      "combined_score": 0.18848828264255746,
      "rank": 232
    },
    {
      "title": "Synergizing Monetization, Orchestration, and Semantics in Computing Continuum",
      "authors": [
        "Chinmaya Kumar Dehury",
        "Lauri Lovén",
        "Praveen Kumar Donta",
        "Ilir Murturi",
        "Schahram Dustdar"
      ],
      "abstract": "Industry demands are growing for hyper-distributed applications that span from the cloud to the edge in domains such as smart manufacturing, transportation, and agriculture. Yet today's solutions struggle to meet these demands due to inherent limitations in scalability, interoperability, and trust. In this article, we introduce HERMES (Heterogeneous Computing Continuum with Resource Monetization, Orchestration, and Semantic) - a novel framework designed to transform connectivity and data utilization across the computing continuum. HERMES establishes an open, seamless, and secure environment where resources, from cloud servers to tiny edge devices, can be orchestrated intelligently, data and services can be monetized in a distributed marketplace, and knowledge is shared through semantic interoperability. By bridging these key facets, HERMES lays a foundation for a new generation of distributed applications that are more efficient, trustworthy, and autonomous.",
      "url": "http://arxiv.org/abs/2512.08288v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.08288v1",
      "citations": null,
      "categories": [
        "cs.DC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.027892990431013197,
      "novelty_score": 0.935064935064935,
      "recency_score": 0.9,
      "relevance_score": 0.188367897129304,
      "bm25_score": 0.0,
      "combined_score": 0.188367897129304,
      "rank": 233
    },
    {
      "title": "Learning Degenerate Manifolds of Frustrated Magnets with Boltzmann Machines",
      "authors": [
        "Jackson C. Glass",
        "Gia-Wei Chern"
      ],
      "abstract": "We show that Restricted Boltzmann Machines (RBMs) provide a flexible generative framework for modeling spin configurations in disordered yet strongly correlated phases of frustrated magnets. As a benchmark, we first demonstrate that an RBM can learn the zero-temperature ground-state manifold of the one-dimensional ANNNI model at its multiphase point, accurately reproducing its characteristic oscillatory and exponentially decaying correlations. We then apply RBMs to kagome spin ice and show that they successfully learn the local ice rules and short-range correlations of the extensively degenerate ice-I manifold. Correlation functions computed from RBM-generated configurations closely match those from direct Monte Carlo simulations. For the partially ordered ice-II phase -- featuring long-range charge order and broken time-reversal symmetry -- accurate modeling requires RBMs with uniform-sign bias fields, mirroring the underlying symmetry breaking. These results highlight the utility of RBMs as generative models for learning constrained and highly frustrated magnetic states.",
      "url": "http://arxiv.org/abs/2511.19879v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.19879v1",
      "citations": null,
      "categories": [
        "cond-mat.str-el",
        "cond-mat.stat-mech",
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.02779196067940661,
      "novelty_score": 0.9329268292682927,
      "recency_score": 0.9,
      "relevance_score": 0.18833758820382202,
      "bm25_score": 0.0,
      "combined_score": 0.18833758820382202,
      "rank": 234
    },
    {
      "title": "GrOMP: Grasped Object Manifold Projection for Multimodal Imitation Learning of Manipulation",
      "authors": [
        "William van den Bogert",
        "Gregory Linkowski",
        "Nima Fazeli"
      ],
      "abstract": "Imitation Learning (IL) holds great potential for learning repetitive manipulation tasks, such as those in industrial assembly. However, its effectiveness is often limited by insufficient trajectory precision due to compounding errors. In this paper, we introduce Grasped Object Manifold Projection (GrOMP), an interactive method that mitigates these errors by constraining a non-rigidly grasped object to a lower-dimensional manifold. GrOMP assumes a precise task in which a manipulator holds an object that may shift within the grasp in an observable manner and must be mated with a grounded part. Crucially, all GrOMP enhancements are learned from the same expert dataset used to train the base IL policy, and are adjusted with an n-arm bandit-based interactive component. We propose a theoretical basis for GrOMP's improvement upon the well-known compounding error bound in IL literature. We demonstrate the framework on four precise assembly tasks using tactile feedback, and note that the approach remains modality-agnostic. Data and videos are available at williamvdb.github.io/GrOMPsite.",
      "url": "http://arxiv.org/abs/2512.03347v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.03347v2",
      "citations": null,
      "categories": [
        "cs.RO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.026987223087471937,
      "novelty_score": 0.925233644859813,
      "recency_score": 0.9,
      "relevance_score": 0.1880961669262416,
      "bm25_score": 0.0,
      "combined_score": 0.1880961669262416,
      "rank": 235
    },
    {
      "title": "Time integration of quantized tensor trains using the interpolative dynamical low-rank approximation",
      "authors": [
        "Erika Ye",
        "Chao Yang"
      ],
      "abstract": "Quantized tensor trains (QTTs) are a low-rank and multiscale framework that allows for efficient approximation and manipulation of multi-dimensional, high resolution data. One area of active research is their use in numerical simulation of hyperbolic systems such as the Navier-Stokes equations and the Vlasov equations. One popular time integration scheme is the dynamical low-rank approximation (DLRA), in which the time integration is constrained to a low-rank manifold. However, until recently, DLRA has typically used orthogonal projectors to project the original dynamical system into a reduced space, which is only well-suited for linear systems. DLRA has also mostly been investigated in the context of non-quantized tensor trains. This work investigates interpolative DLRA schemes in which the low-rank manifold is constructed from aptly chosen interpolation points and interpolating polynomials, in the context of QTTs. Through various examples, its performance is compared to its orthogonal counterpart. This work demonstrates how interpolative DLRA is suitable for nonlinear systems and time integrators requiring nonlinear element-wise operations, such as upwind time integration schemes.",
      "url": "http://arxiv.org/abs/2512.15703v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.15703v1",
      "citations": null,
      "categories": [
        "math.NA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.024961916054590916,
      "novelty_score": 0.9459041731066461,
      "recency_score": 0.9,
      "relevance_score": 0.1874885748163773,
      "bm25_score": 0.0,
      "combined_score": 0.1874885748163773,
      "rank": 236
    },
    {
      "title": "Attention Is Not What You Need",
      "authors": [
        "Zhang Chong"
      ],
      "abstract": "We revisit a basic question in sequence modeling: is explicit self-attention actually necessary for strong performance and reasoning? We argue that standard multi-head attention is best seen as a form of tensor lifting: hidden vectors are mapped into a high-dimensional space of pairwise interactions, and learning proceeds by constraining this lifted tensor through gradient descent. This mechanism is extremely expressive but mathematically opaque, because after many layers it becomes very hard to describe the model with a small family of explicit invariants.\n  To explore an alternative, we propose an attention-free architecture based on Grassmann flows. Instead of forming an L by L attention matrix, our Causal Grassmann layer (i) linearly reduces token states, (ii) encodes local token pairs as two-dimensional subspaces on a Grassmann manifold via Plucker coordinates, and (iii) fuses these geometric features back into the hidden states through gated mixing. Information therefore propagates by controlled deformations of low-rank subspaces over multi-scale local windows, so the core computation lives on a finite-dimensional manifold rather than in an unstructured tensor space.\n  On the Wikitext-2 language modeling benchmark, purely Grassmann-based models with 13 to 18 million parameters achieve validation perplexities within about 10 to 15 percent of size-matched Transformers. On the SNLI natural language inference task, a Grassmann-Plucker head on top of DistilBERT slightly outperforms a Transformer head, with best validation and test accuracies of 0.8550 and 0.8538 compared to 0.8545 and 0.8511. We analyze the complexity of Grassmann mixing, show linear scaling in sequence length for fixed rank, and argue that such manifold-based designs offer a more structured route toward geometric and invariant-based interpretations of neural reasoning.",
      "url": "http://arxiv.org/abs/2512.19428v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.19428v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI",
        "math.AG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.02303977183689576,
      "novelty_score": 0.9924324324324324,
      "recency_score": 0.9,
      "relevance_score": 0.18691193155106875,
      "bm25_score": 0.0,
      "combined_score": 0.18691193155106875,
      "rank": 237
    },
    {
      "title": "Hybrid twinning using PBDW and DeepONet for the effective state estimation and prediction on partially known systems",
      "authors": [
        "Stiven Briand Massala",
        "Ludovic Chamoin",
        "Massimo Picca Ciamarra"
      ],
      "abstract": "The accurate estimation of the state of complex uncertain physical systems requires reconciling theoretical models, with inherent imperfections, with noisy experimental data. In this work, we propose an effective hybrid approach that combines physics-based modeling with data-driven learning to enhance state estimation and further prediction. Our method builds upon the Parameterized Background Data-Weak (PBDW) framework, which naturally integrates a reduced-order representation of the best-available model with measurement data to account for both anticipated and unanticipated uncertainties. To address model discrepancies not captured by the reduced-order space, and learn the structure of model deviation, we incorporate a Deep Operator Network (DeepONet) constrained to be an orthogonal complement of the best-knowledge manifold. This ensures that the learned correction targets only the unknown components of model bias, preserving the interpretability and fidelity of the physical model. An optimal sensor placement strategy is also investigated to maximize information gained from measurements. We validate the proposed approach on a representative problem involving the Helmholtz equation under various sources of modeling error, including those arising from boundary conditions and source terms.",
      "url": "http://arxiv.org/abs/2512.11834v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.11834v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.021877795417468367,
      "novelty_score": 0.9168135354247444,
      "recency_score": 0.9,
      "relevance_score": 0.18656333862524052,
      "bm25_score": 0.0,
      "combined_score": 0.18656333862524052,
      "rank": 238
    },
    {
      "title": "Differentiable Inverse Modeling with Physics-Constrained Latent Diffusion for Heterogeneous Subsurface Parameter Fields",
      "authors": [
        "Zihan Lin",
        "QiZhi He"
      ],
      "abstract": "We present a latent diffusion-based differentiable inversion method (LD-DIM) for PDE-constrained inverse problems involving high-dimensional spatially distributed coefficients. LD-DIM couples a pretrained latent diffusion prior with an end-to-end differentiable numerical solver to reconstruct unknown heterogeneous parameter fields in a low-dimensional nonlinear manifold, improving numerical conditioning and enabling stable gradient-based optimization under sparse observations. The proposed framework integrates a latent diffusion model (LDM), trained in a compact latent space, with a differentiable finite-volume discretization of the forward PDE. Sensitivities are propagated through the discretization using adjoint-based gradients combined with reverse-mode automatic differentiation. Inversion is performed directly in latent space, which implicitly suppresses ill-conditioned degrees of freedom while preserving dominant structural modes, including sharp material interfaces. The effectiveness of LD-DIM is demonstrated using a representative inverse problem for flow in porous media, where heterogeneous conductivity fields are reconstructed from spatially sparse hydraulic head measurements. Numerical experiments assess convergence behavior and reconstruction quality for both Gaussian random fields and bimaterial coefficient distributions. The results show that LD-DIM achieves consistently improved numerical stability and reconstruction accuracy of both parameter fields and corresponding PDE solutions compared with physics-informed neural networks (PINNs) and physics-embedded variational autoencoder (VAE) baselines, while maintaining sharp discontinuities and reducing sensitivity to initialization.",
      "url": "http://arxiv.org/abs/2512.22421v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.22421v1",
      "citations": null,
      "categories": [
        "math.NA",
        "cs.LG",
        "physics.geo-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.02185485712981061,
      "novelty_score": 0.945417095777549,
      "recency_score": 0.9,
      "relevance_score": 0.1865564571389432,
      "bm25_score": 0.0,
      "combined_score": 0.1865564571389432,
      "rank": 239
    },
    {
      "title": "An Empirical Study of Sampling Hyperparameters in Diffusion-Based Super-Resolution",
      "authors": [
        "Yudhistira Arief Wibowo"
      ],
      "abstract": "Diffusion models have shown strong potential for solving inverse problems such as single-image super-resolution, where a high-resolution image is recovered from a low-resolution observation using a pretrained unconditional prior. Conditioning methods, including Diffusion Posterior Sampling (DPS) and Manifold Constrained Gradient (MCG), can substantially improve reconstruction quality, but they introduce additional hyperparameters that require careful tuning. In this work, we conduct an empirical ablation study on FFHQ super-resolution to identify the dominant factors affecting performance when applying conditioning to pretrained diffusion models, and show that the conditioning step size has a significantly greater impact than the diffusion step count, with step sizes in the range of [2.0, 3.0] yielding the best overall performance in our experiments.",
      "url": "http://arxiv.org/abs/2512.17675v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.17675v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.02147097325216803,
      "novelty_score": 0.9392905866302864,
      "recency_score": 0.9,
      "relevance_score": 0.18644129197565043,
      "bm25_score": 0.0,
      "combined_score": 0.18644129197565043,
      "rank": 240
    },
    {
      "title": "Situationally Sensitive Path Planning",
      "authors": [
        "Paul M. Torrens",
        "Ryan Kim",
        "Kaishuu Shinozaki-Conefrey"
      ],
      "abstract": "We examine how site-based path planning algorithms for enclosed spaces can be enhanced with situational detail. Addressing this question has led to value propositions in facility design, where there is often a call to match, map, and merge infrastructure considerations and configurations with potential implications for individual, group, and crowd flow through enclosed spaces. Responding to this question also invokes computational propositions, as facility design software is often computationally conservative with few resources devoted to simulation. We show that situational factors—the peculiarities and momentarily fleeting shifts in an individualized context that embody people in their movement through spaces—can be embedded into traditional, computationally lean path planning heuristics in ways that are actionable in widely used facility design software. We achieve this with algorithmic expansion of well-known planning algorithms using node-based architectures that permit the inclusion detail if, when, and where needed in a hyper-localized situational context that nests within site considerations. We demonstrate a proof of concept for use in the popular Unity 3D modeling platform, showing that situationally sensitive path planning can be achieved during the simulation run time of prototypical design scenarios for enclosed spaces with moving individuals, groups, and crowds.",
      "url": "https://openalex.org/W4411690641",
      "year": 2025,
      "venue": "Algorithms",
      "source": "openalex",
      "doi": "10.3390/a18070388",
      "pdf_url": "https://www.mdpi.com/1999-4893/18/7/388/pdf?version=1751024551",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.021315901470820443,
      "novelty_score": 0.9870967741935485,
      "recency_score": 0.9,
      "relevance_score": 0.18639477044124617,
      "bm25_score": 0.0,
      "combined_score": 0.18639477044124617,
      "rank": 241
    },
    {
      "title": "Dark Matter Induced Nucleon Decay Through the Neutron Portal",
      "authors": [
        "Nicole F. Bell",
        "Peter Cox",
        "Jayden L. Newstead",
        "Michael B. G. Verde"
      ],
      "abstract": "The neutron portal operator provides a theoretically motivated connection between the visible and dark sectors and features in several well-studied asymmetric dark matter models. This operator leads to dark matter induced nucleon decays that mimic the experimental signature of \"ordinary\" nucleon decays. In this work, we reinterpret Super-Kamiokande nucleon decay searches for $n \\rightarrow π^0 ν$ and $p \\rightarrow π^+ ν$ to constrain dark matter induced nucleon decays. For GeV-scale dark matter, we obtain lower bounds of $\\mathcal{O}(1~\\rm{TeV})$ on the scale of the effective neutron portal operator. We also discuss the prospects for future searches at Hyper-Kamiokande and highlight the importance of a dedicated experimental analysis with reduced systematic uncertainties.",
      "url": "http://arxiv.org/abs/2511.18722v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.18722v1",
      "citations": null,
      "categories": [
        "hep-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.02127506296818516,
      "novelty_score": 0.966994382022472,
      "recency_score": 0.9,
      "relevance_score": 0.18638251889045557,
      "bm25_score": 0.0,
      "combined_score": 0.18638251889045557,
      "rank": 242
    },
    {
      "title": "Secure Analog Beamforming for Multi-user MISO Systems with Movable Antennas",
      "authors": [
        "Weijie Xiong",
        "Jingran Lin",
        "Kai Zhong",
        "Liu Yang",
        "Hongli Liu",
        "Qiang Li",
        "Cunhua Pan"
      ],
      "abstract": "Movable antennas (MAs) represent a novel approach that enables flexible adjustments to antenna positions, effectively altering the channel environment and thereby enhancing the performance of wireless communication systems. However, conventional MA implementations often adopt fully digital beamforming (FDB), which requires a dedicated RF chain for each antenna. This requirement significantly increase hardware costs, making such systems impractical for multi-antenna deployments. To address this, hardware-efficient analog beamforming (AB) offers a cost-effective alternative. This paper investigates the physical layer security (PLS) in an MA-enabled multiple-input single-output (MISO) communication system with an emphasis on AB. In this scenario, an MA-enabled transmitter with AB broadcasts common confidential information to a group of legitimate receivers, while a number of eavesdroppers overhear the transmission and attempt to intercept the information. Our objective is to maximize the multicast secrecy rate (MSR) by jointly optimizing the phase shifts of the AB and the positions of the MAs, subject to constraints on the movement area of the MAs and the constant modulus (CM) property of the analog phase shifters. This MSR maximization problem is highly challenging, as we have formally proven it to be NP-hard. To solve it efficiently, we propose a penalty constrained product manifold (PCPM) framework. Specifically, we first reformulate the position constraints as a penalty function, enabling unconstrained optimization on a product manifold space (PMS), and then propose a parallel conjugate gradient descent algorithm to efficiently update the variables. Simulation results demonstrate that MA-enabled systems with AB can achieve a well-balanced performance in terms of MSR and hardware costs.",
      "url": "http://arxiv.org/abs/2511.19360v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.19360v1",
      "citations": null,
      "categories": [
        "eess.SP"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.021254459327523602,
      "novelty_score": 0.9444444444444444,
      "recency_score": 0.9,
      "relevance_score": 0.1863763377982571,
      "bm25_score": 0.0,
      "combined_score": 0.1863763377982571,
      "rank": 243
    },
    {
      "title": "Chiral Magnetic Effect induced Spectator Process for Leptogenesis",
      "authors": [
        "Wei Chao"
      ],
      "abstract": "Conventional Leptogenesis mechanism, which provides compelling explanation to the origin of the baryon asymmetry of the universe (BAU), assumes the absence of hypermagnetic field in the early universe, thereby disregard the implications of hyper gauge field helicity, that have been thoroughly studied in the magnetogenesis mechanism. In this paper, we address impacts of a general U(1) gauge field on Leptogenesis by deriving equation of motions for the helicity and the energy density of a general magnetic field, to which the chiral magnetic effect (CME) is identified as essential, and studying their effects on the evolution chiral asymmetries. Notably, CME in the $U(1)_{\\mathbf{L}_i-\\mathbf{L}_j}$ framework, where $\\mathbf{L}_{i,j}$ means specific lepton flavor, explicitly breaks the total lepton number and provides an efficient spectator process, that can wash out pre-existing lepton asymmetries. This establishes a natural connection to the wash-in Leptogenesis paradigm. We demonstrate that this spectator effect enables the generation of the BAU, eliminating the need for both an initial $\\mathbf{B}-\\mathbf{L}$ charge and primordial helicity.",
      "url": "http://arxiv.org/abs/2511.13051v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.13051v1",
      "citations": null,
      "categories": [
        "hep-ph",
        "astro-ph.CO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.02087593804499658,
      "novelty_score": 0.9488372093023255,
      "recency_score": 0.9,
      "relevance_score": 0.186262781413499,
      "bm25_score": 0.0,
      "combined_score": 0.186262781413499,
      "rank": 244
    },
    {
      "title": "Bhargava Cube--Inspired Quadratic Regularization for Structured Neural Embeddings",
      "authors": [
        "S Sairam",
        "Prateek P Kulkarni"
      ],
      "abstract": "We present a novel approach to neural representation learning that incorporates algebraic constraints inspired by Bhargava cubes from number theory. Traditional deep learning methods learn representations in unstructured latent spaces lacking interpretability and mathematical consistency. Our framework maps input data to constrained 3-dimensional latent spaces where embeddings are regularized to satisfy learned quadratic relationships derived from Bhargava's combinatorial structures. The architecture employs a differentiable auxiliary loss function operating independently of classification objectives, guiding models toward mathematically structured representations. We evaluate on MNIST, achieving 99.46% accuracy while producing interpretable 3D embeddings that naturally cluster by digit class and satisfy learned quadratic constraints. Unlike existing manifold learning approaches requiring explicit geometric supervision, our method imposes weak algebraic priors through differentiable constraints, ensuring compatibility with standard optimization. This represents the first application of number-theoretic constructs to neural representation learning, establishing a foundation for incorporating structured mathematical priors in neural networks.",
      "url": "http://arxiv.org/abs/2512.11392v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.11392v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.02052779337216456,
      "novelty_score": 0.9422632794457275,
      "recency_score": 0.9,
      "relevance_score": 0.18615833801164938,
      "bm25_score": 0.0,
      "combined_score": 0.18615833801164938,
      "rank": 245
    },
    {
      "title": "Transmit Weights, Not Features: Orthogonal-Basis Aided Wireless Point-Cloud Transmission",
      "authors": [
        "Junlin Chang",
        "Yubo Han",
        "Hnag Yue",
        "John S Thompson",
        "Rongke Liu"
      ],
      "abstract": "The widespread adoption of depth sensors has substantially lowered the barrier to point-cloud acquisition. This letter proposes a semantic wireless transmission framework for three dimension (3D) point clouds built on Deep Joint Source - Channel Coding (DeepJSCC). Instead of sending raw features, the transmitter predicts combination weights over a receiver-side semantic orthogonal feature pool, enabling compact representations and robust reconstruction. A folding-based decoder deforms a 2D grid into 3D, enforcing manifold continuity while preserving geometric fidelity. Trained with Chamfer Distance (CD) and an orthogonality regularizer, the system is evaluated on ModelNet40 across varying Signal-to-Noise Ratios (SNRs) and bandwidths. Results show performance on par with SEmantic Point cloud Transmission (SEPT) at high bandwidth and clear gains in bandwidth-constrained regimes, with consistent improvements in both Peak Signal-to-Noise Ratio (PSNR) and CD. Ablation experiments confirm the benefits of orthogonalization and the folding prior.",
      "url": "http://arxiv.org/abs/2512.03819v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.03819v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.019933857599098173,
      "novelty_score": 0.9920749279538905,
      "recency_score": 0.9,
      "relevance_score": 0.18598015727972947,
      "bm25_score": 0.0,
      "combined_score": 0.18598015727972947,
      "rank": 246
    },
    {
      "title": "Reconstructing Multi-Scale Physical Fields from Extremely Sparse Measurements with an Autoencoder-Diffusion Cascade",
      "authors": [
        "Letian Yi",
        "Tingpeng Zhang",
        "Mingyuan Zhou",
        "Guannan Wang",
        "Quanke Su",
        "Zhilu Lai"
      ],
      "abstract": "Reconstructing full fields from extremely sparse and random measurements is a longstanding ill-posed inverse problem. A powerful framework for addressing such challenges is hierarchical probabilistic modeling, where uncertainty is represented by intermediate variables and resolved through marginalization during inference. Inspired by this principle, we propose Cascaded Sensing (Cas-Sensing), a hierarchical reconstruction framework that integrates an autoencoder-diffusion cascade. First, a neural operator-based functional autoencoder reconstructs the dominant structures of the original field - including large-scale components and geometric boundaries - from arbitrary sparse inputs, serving as an intermediate variable. Then, a conditional diffusion model, trained with a mask-cascade strategy, generates fine-scale details conditioned on these large-scale structures. To further enhance fidelity, measurement consistency is enforced via the manifold constrained gradient based on Bayesian posterior sampling during the generation process. This cascaded pipeline substantially alleviates ill-posedness, delivering accurate and robust reconstructions. Experiments on both simulation and real-world datasets demonstrate that Cas-Sensing generalizes well across varying sensor configurations and geometric boundaries, making it a promising tool for practical deployment in scientific and engineering applications.",
      "url": "http://arxiv.org/abs/2512.01572v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.01572v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI",
        "physics.app-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.01858715434518546,
      "novelty_score": 0.9765957446808511,
      "recency_score": 0.9,
      "relevance_score": 0.18557614630355565,
      "bm25_score": 0.0,
      "combined_score": 0.18557614630355565,
      "rank": 247
    },
    {
      "title": "Breaking Symmetry-Induced Degeneracy in Multi-Agent Ergodic Coverage via Stochastic Spectral Control",
      "authors": [
        "Kooktae Lee",
        "Julian Martinez"
      ],
      "abstract": "Multi-agent ergodic coverage via Spectral Multiscale Coverage (SMC) provides a principled framework for driving a team of agents so that their collective time-averaged trajectories match a prescribed spatial distribution. While classical SMC has demonstrated empirical success, it can suffer from gradient cancellation, particularly when agents are initialized near symmetry points of the target distribution, leading to undesirable behaviors such as stalling or motion constrained along symmetry axes. In this work, we rigorously characterize the initial conditions and symmetry-induced invariant manifolds that give rise to such directional degeneracy in first-order agent dynamics. To address this, we introduce a stochastic perturbation combined with a contraction term and prove that the resulting dynamics ensure almost-sure escape from zero-gradient manifolds while maintaining mean-square boundedness of agent trajectories. Simulations on symmetric multi-modal reference distributions demonstrate that the proposed stochastic SMC effectively mitigates transient stalling and axis-constrained motion, while ensuring that all agent trajectories remain bounded within the domain.",
      "url": "http://arxiv.org/abs/2512.23158v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.23158v1",
      "citations": null,
      "categories": [
        "eess.SY",
        "cs.RO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.018331331749927373,
      "novelty_score": 0.9722703639514731,
      "recency_score": 0.9,
      "relevance_score": 0.18549939952497824,
      "bm25_score": 0.0,
      "combined_score": 0.18549939952497824,
      "rank": 248
    },
    {
      "title": "Complexity guarantees and polling strategies for Riemannian direct-search methods",
      "authors": [
        "Bastien Cavarretta",
        "Florentin Goyens",
        "Clément W. Royer",
        "Florian Yger"
      ],
      "abstract": "Direct-search algorithms are derivative-free optimization techniques that operate by polling the variable space along specific directions forming positive spanning sets (PSSs). When the problem variables are constrained to lie on a Riemannian manifold, polling must be performed along tangent directions. Although Riemannian variants of direct search have already been proposed and endowed with asymptotic guarantees, a proper generalization of PSSs on manifolds remains to be investigated. In particular, a measure of quality for those PSSs is required to obtain complexity bounds for direct search.\n  In this paper, we derive complexity guarantees for a class of Riemannian direct-search techniques, and study two ways of generating positive spanning sets in tangent spaces. We pay particular attention the unit hypersphere case, for which we establish that generating directions directly within the tangent space leads to better complexity properties than projecting PSSs from the ambient space onto the tangent space. Our numerical experiments highlight the impact of dimension and codimension in more general settings.",
      "url": "http://arxiv.org/abs/2511.15360v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.15360v1",
      "citations": null,
      "categories": [
        "math.OC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.01768488030931645,
      "novelty_score": 0.9198396793587174,
      "recency_score": 0.9,
      "relevance_score": 0.18530546409279497,
      "bm25_score": 0.0,
      "combined_score": 0.18530546409279497,
      "rank": 249
    },
    {
      "title": "Neural Network Optimal Power Flow via Energy Gradient Flow and Unified Dynamics",
      "authors": [
        "Xuezhi Liu"
      ],
      "abstract": "Optimal Power Flow (OPF) is a core optimization problem in power system operation and planning, aiming to minimize generation costs while satisfying physical constraints such as power flow equations, generator limits, and voltage limits. Traditional OPF solving methods typically employ iterative optimization algorithms (such as interior point methods, sequential quadratic programming, etc.), with limitations including low computational efficiency, initial value sensitivity, and low batch computation efficiency. Most existing deep learning-based OPF methods rely on supervised learning, requiring pre-solving large numbers of cases, and have difficulty guaranteeing physical consistency. This paper proposes an Optimal Power Flow solving method based on neural network dynamics and energy gradient flow, transforming OPF problems into energy minimization problems. By constructing an energy function to measure the degree of deviation from the constraint manifold, and guiding networks to learn optimal solutions that simultaneously satisfy power flow constraints and minimize costs through gradient flow. Neural networks are trained unsupervised by directly minimizing physical residuals, requiring no labeled data, achieving true \"end-to-end\" physics-constrained learning.",
      "url": "http://arxiv.org/abs/2512.01219v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.01219v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.016698167722528984,
      "novelty_score": 0.9459041731066461,
      "recency_score": 0.9,
      "relevance_score": 0.18500945031675872,
      "bm25_score": 0.0,
      "combined_score": 0.18500945031675872,
      "rank": 250
    },
    {
      "title": "Geometry and quantum brachistochrone analysis of multiple entangled spin-1/2 particles under all-range Ising interaction",
      "authors": [
        "B. Amghar",
        "M. Yachi",
        "M. Amghar",
        "M. Almousa",
        "A. A. Abd El-Latif",
        "A. Slaoui"
      ],
      "abstract": "We present a unified geometric and dynamical framework for a physical system consisting of $n$ spin-$1/2$ particles with all-range Ising interaction. Using the Fubini-Study formalism, we derive the metric tensor of the associated quantum state manifold and compute the corresponding Riemann curvature. Our analysis reveals that the system evolves over a smooth, compact, two-dimensional manifold with spherical topology and a dumbbell-like structure shaped by collective spin interactions. We further investigate the influence of the geometry and topology of the resulting state space on the behavior of geometric and topological phases acquired by the system. We explore how this curvature constrains the system's dynamical behavior, including its evolution speed and Fubini-Study distance between the quantum states. Within this geometric framework, we address the quantum brachistochrone problem and derive the minimal time required for optimal evolution, a result useful for time-efficient quantum circuit design. Subsequently, we explore the role of entanglement in shaping the state space geometry, modulating geometric phase, and controlling evolution speed and brachistochrone time. Our results reveal that entanglement enhances dynamics up to a critical threshold, beyond which geometric constraints begin to hinder evolution. Moreover, entanglement induces critical shifts in the geometric phase, making it a sensitive indicator of entanglement levels and a practical tool for steering quantum evolution. This approach offers valuable guidance for developing quantum technologies that require time-efficient control strategies rooted in the geometry of quantum state space.",
      "url": "http://arxiv.org/abs/2512.21400v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": "10.1038/s41598-025-32484-y",
      "pdf_url": "https://arxiv.org/pdf/2512.21400v1",
      "citations": null,
      "categories": [
        "quant-ph",
        "math-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.01600642132967346,
      "novelty_score": 0.9444444444444444,
      "recency_score": 0.9,
      "relevance_score": 0.18480192639890206,
      "bm25_score": 0.0,
      "combined_score": 0.18480192639890206,
      "rank": 251
    },
    {
      "title": "The Homological Brain: Parity Principle and Amortized Inference",
      "authors": [
        "Xin Li"
      ],
      "abstract": "Biological intelligence emerges from substrates that are slow, noisy, and energetically constrained, yet it performs rapid and coherent inference in open-ended environments. Classical computational theories, built around vector-space transformations and instantaneous error minimization, struggle to reconcile the slow timescale of synaptic plasticity with the fast timescale of perceptual synthesis. We propose a unifying framework based on algebraic topology, the Homological Brain, in which neural computation is understood as the construction and navigation of topological structure. Central to this view is the Parity Principle, a homological partition between even-dimensional scaffolds encoding stable content ($Φ$) and odd-dimensional flows encoding dynamic context ($Ψ$). Transient contextual flows are resolved through a three-stage topological trinity transformation: Search (open-chain exploration), Closure (topological cycle formation), and Condensation (collapse of validated flows into new scaffold). This process converts high-complexity recursive search (formally modeled by Savitch's Theorem in NPSPACE) into low-complexity navigation over a learned manifold (analogous to memoized Dynamic Programming in P). In this framework, topological condensation is the mechanism that transforms a ``search problem'' into a ``navigation task'', allowing the brain to amortize past inference and achieve rapid perceptual integration. This perspective unifies the Wake-Sleep cycle, episodic-to-semantic consolidation, and dual-process theories (System 1-vs-System 2), revealing the brain as a homology engine that minimizes topological complexity to transmute high-entropy sensory flux into low-entropy, invariant cognitive structure.",
      "url": "http://arxiv.org/abs/2512.10976v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.10976v1",
      "citations": null,
      "categories": [
        "q-bio.NC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.01577177323563017,
      "novelty_score": 0.9315068493150686,
      "recency_score": 0.9,
      "relevance_score": 0.18473153197068906,
      "bm25_score": 0.0,
      "combined_score": 0.18473153197068906,
      "rank": 252
    },
    {
      "title": "Training-Free Diffusion Priors for Text-to-Image Generation via Optimization-based Visual Inversion",
      "authors": [
        "Samuele Dell'Erba",
        "Andrew D. Bagdanov"
      ],
      "abstract": "Diffusion models have established the state-of-the-art in text-to-image generation, but their performance often relies on a diffusion prior network to translate text embeddings into the visual manifold for easier decoding. These priors are computationally expensive and require extensive training on massive datasets. In this work, we challenge the necessity of a trained prior at all by employing Optimization-based Visual Inversion (OVI), a training-free and zero-shot alternative, to replace the need for a prior. OVI initializes a latent visual representation from random pseudo-tokens and iteratively optimizes it to maximize the cosine similarity with the input textual prompt embedding. We further propose two novel constraints, a Mahalanobis-based and a Nearest-Neighbor loss, to regularize the OVI optimization process toward the distribution of realistic images. Our experiments, conducted on Kandinsky 2.2, show that OVI can serve as an alternative to traditional priors. More importantly, our analysis reveals a critical flaw in current evaluation benchmarks like T2I-CompBench++, where simply using the text embedding as a prior achieves surprisingly high scores, despite lower perceptual quality. Our constrained OVI methods improve visual fidelity over this baseline, with the Nearest-Neighbor approach proving particularly effective. It achieves quantitative scores comparable to or higher than the state-of-the-art data-efficient prior, underscoring the potential of optimization-based strategies as viable, training-free alternatives to traditional priors. The code will be publicly available upon acceptance.",
      "url": "http://arxiv.org/abs/2511.20821v3",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.20821v3",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.AI",
        "cs.CL",
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.015538472768811935,
      "novelty_score": 0.9467821782178217,
      "recency_score": 0.9,
      "relevance_score": 0.1846615418306436,
      "bm25_score": 0.0,
      "combined_score": 0.1846615418306436,
      "rank": 253
    },
    {
      "title": "Connection Between Dwarf Galaxies and Globular Clusters: Insights from the Perseus Cluster Using Subaru Imaging and Keck Spectroscopy",
      "authors": [
        "Yimeng Tang",
        "Aaron J. Romanowsky",
        "Song Huang",
        "Nobuhiro Okabe",
        "Jean P. Brodie",
        "Kevin A. Bundy",
        "Maria Luisa Buzzo",
        "Timothy Carleton",
        "Anna Ferré-Mateu",
        "Duncan A. Forbes"
      ],
      "abstract": "We present a systematic study of 189 dwarf galaxies and their globular cluster (GC) systems in the Perseus cluster, based on deep Subaru Hyper Suprime-Cam imaging and Keck spectroscopy, supplemented by literature data. This constitutes the largest sample of dwarfs in a single galaxy cluster to date with simultaneous deep imaging, spectroscopic coverage, and GC measurements, while uniquely spanning a broad and continuous range of galaxy properties. We find an anti-correlation between GC specific mass and galaxy stellar mass for dwarfs in Perseus similar to observations in other clusters. At fixed stellar mass, dwarfs with lower surface brightness or larger effective radius tend to be more GC-rich -- suggesting either high GC formation efficiency in an earlier compact-galaxy phase, or less efficient GC disruption. The correlation between GC richness and axis ratio in Perseus is weaker than in other environments. We find some connection between GC richness and infall time, but not with the clear correlations found in Virgo, Coma, and cosmological simulations. More complete observations are needed to test for cluster-to-cluster variations in galaxy and GC evolutionary histories. This work demonstrates the potential of new wide-field imaging and spectroscopy surveys for understanding GCs and dwarf galaxies, and highlights the need for further work in theoretical modeling.",
      "url": "http://arxiv.org/abs/2512.11070v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.11070v1",
      "citations": null,
      "categories": [
        "astro-ph.GA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.015270125699603227,
      "novelty_score": 0.9457417582417582,
      "recency_score": 0.9,
      "relevance_score": 0.184581037709881,
      "bm25_score": 0.0,
      "combined_score": 0.184581037709881,
      "rank": 254
    },
    {
      "title": "Lotus-2: Advancing Geometric Dense Prediction with Powerful Image Generative Model",
      "authors": [
        "Jing He",
        "Haodong Li",
        "Mingzhi Sheng",
        "Ying-Cong Chen"
      ],
      "abstract": "Recovering pixel-wise geometric properties from a single image is fundamentally ill-posed due to appearance ambiguity and non-injective mappings between 2D observations and 3D structures. While discriminative regression models achieve strong performance through large-scale supervision, their success is bounded by the scale, quality and diversity of available data and limited physical reasoning. Recent diffusion models exhibit powerful world priors that encode geometry and semantics learned from massive image-text data, yet directly reusing their stochastic generative formulation is suboptimal for deterministic geometric inference: the former is optimized for diverse and high-fidelity image generation, whereas the latter requires stable and accurate predictions. In this work, we propose Lotus-2, a two-stage deterministic framework for stable, accurate and fine-grained geometric dense prediction, aiming to provide an optimal adaption protocol to fully exploit the pre-trained generative priors. Specifically, in the first stage, the core predictor employs a single-step deterministic formulation with a clean-data objective and a lightweight local continuity module (LCM) to generate globally coherent structures without grid artifacts. In the second stage, the detail sharpener performs a constrained multi-step rectified-flow refinement within the manifold defined by the core predictor, enhancing fine-grained geometry through noise-free deterministic flow matching. Using only 59K training samples, less than 1% of existing large-scale datasets, Lotus-2 establishes new state-of-the-art results in monocular depth estimation and highly competitive surface normal prediction. These results demonstrate that diffusion models can serve as deterministic world priors, enabling high-quality geometric reasoning beyond traditional discriminative and generative paradigms.",
      "url": "http://arxiv.org/abs/2512.01030v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.01030v2",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.015245102658659573,
      "novelty_score": 0.9677419354838709,
      "recency_score": 0.9,
      "relevance_score": 0.1845735307975979,
      "bm25_score": 0.0,
      "combined_score": 0.1845735307975979,
      "rank": 255
    },
    {
      "title": "Transversal Gates in Nonadditive Quantum Codes",
      "authors": [
        "Chao Zhang",
        "Zipeng Wu",
        "Shilin Huang",
        "Bei Zeng"
      ],
      "abstract": "Transversal gates play a crucial role in suppressing error propagation in fault-tolerant quantum computation, yet they are intrinsically constrained: any nontrivial code encoding a single logical qubit admits only a finite subgroup of $\\mathrm{SU}(2)$ as its transversal operations. We introduce a systematic framework for searching codes with specified transversal groups by parametrizing their logical subspaces on the Stiefel manifold and minimizing a composite loss that enforces both the Knill-Laflamme conditions and a target transversal-group structure. Applying this method, we uncover a new $((6,2,3))$ code admitting a transversal $Z\\bigl(\\tfrac{2\\pi}{5}\\bigr)$ gate (transversal group $\\mathrm{C}_{10}$), the smallest known distance $3$ code supporting non-Clifford transversal gates, as well as several new $((7,2,3))$ codes realizing the binary icosahedral group $2I$. We further propose the \\emph{Subset-Sum-Linear-Programming} (SS-LP) construction for codes with transversal \\emph{diagonal} gates, which dramatically shrinks the search space by reducing to integer partitions subject to linear constraints. In a more constrained form, the method also applies directly to the binary-dihedral groups $\\mathrm{BD}_{2m}$. Specializing to $n=7$, the SS-LP method yields codes for all $\\mathrm{BD}_{2m}$ with $2m\\le 36$, including the first $((7,2,3))$ examples supporting transversal $T$ gate ($\\mathrm{BD}_{16}$) and $\\sqrt{T}$ gate ($\\mathrm{BD}_{32}$), improving on the previous smallest examples $((11,2,3))$ and $((19,2,3))$. Extending the SS-LP approach to $((8,2,3))$, we construct new codes for $2m>36$, including one supporting a transversal $T^{1/4}$ gate ($\\mathrm{BD}_{64}$). These results reveal a far richer landscape of nonadditive codes than previously recognized and underscore a deeper connection between quantum error correction and the algebraic constraints on transversal gate groups.",
      "url": "https://www.semanticscholar.org/paper/858abd26e4f99dda2468663d810267fde539c457",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.014829893830740664,
      "novelty_score": 0.9542619542619543,
      "recency_score": 0.9,
      "relevance_score": 0.1844489681492222,
      "bm25_score": 0.0,
      "combined_score": 0.1844489681492222,
      "rank": 256
    },
    {
      "title": "Geometric Control of Mechanical Systems with Symmetries Based on Sliding Modes",
      "authors": [
        "Eduardo Espíndola",
        "Yu Tang"
      ],
      "abstract": "In this paper, we propose a framework for designing sliding mode controllers for a class of mechanical systems with symmetry, both unconstrained and constrained, that evolve on principal fiber bundles. Control laws are developed based on the reduced motion equations by exploring symmetries, leading to a sliding mode control strategy where the reaching stage is executed on the base space, and the sliding stage is performed on the structure group. Thus, design complexity is reduced, and difficult choices for coordinate representations when working with a particular Lie group are avoided. For this purpose, a sliding subgroup is constructed on the structure group based on a kinematic controller, and the sliding variable will converge to the identity of the state manifold upon reaching the sliding subgroup. A reaching law based on a general sliding vector field is then designed on the base space using the local form of the mechanical connection to drive the sliding variable to the sliding subgroup, and its time evolution is given according to the appropriate covariant derivative. Almost global asymptotic stability and local exponential stability are demonstrated using a Lyapunov analysis. We apply the results to a fully actuated system (a rigid spacecraft actuated by reaction wheels) and a subactuated nonholonomic system (unicycle mobile robot actuated by wheels), which is also simulated for illustration.",
      "url": "https://www.semanticscholar.org/paper/358576ff413689af42082670164c424f7898735d",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2509.01985",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.014194903008989195,
      "novelty_score": 0.9349999999999999,
      "recency_score": 0.9,
      "relevance_score": 0.18425847090269679,
      "bm25_score": 0.0,
      "combined_score": 0.18425847090269679,
      "rank": 257
    },
    {
      "title": "Kinetic-Mamba: Mamba-Assisted Predictions of Stiff Chemical Kinetics",
      "authors": [
        "Additi Pandey",
        "Liang Wei",
        "Hessam Babaee",
        "George Em Karniadakis"
      ],
      "abstract": "Accurate chemical kinetics modeling is essential for combustion simulations, as it governs the evolution of complex reaction pathways and thermochemical states. In this work, we introduce Kinetic-Mamba, a Mamba-based neural operator framework that integrates the expressive power of neural operators with the efficient temporal modeling capabilities of Mamba architectures. The framework comprises three complementary models: (i) a standalone Mamba model that predicts the time evolution of thermochemical state variables from given initial conditions; (ii) a constrained Mamba model that enforces mass conservation while learning the state dynamics; and (iii) a regime-informed architecture employing two standalone Mamba models to capture dynamics across temperature-dependent regimes. We additionally develop a latent Kinetic-Mamba variant that evolves dynamics in a reduced latent space and reconstructs the full state on the physical manifold. We evaluate the accuracy and robustness of Kinetic-Mamba using both time-decomposition and recursive-prediction strategies. We further assess the extrapolation capabilities of the model on varied out-of-distribution datasets. Computational experiments on Syngas and GRI-Mech 3.0 reaction mechanisms demonstrate that our framework achieves high fidelity in predicting complex kinetic behavior using only the initial conditions of the state variables.",
      "url": "http://arxiv.org/abs/2512.14471v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.14471v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.014150406526977882,
      "novelty_score": 0.952,
      "recency_score": 0.9,
      "relevance_score": 0.1842451219580934,
      "bm25_score": 0.0,
      "combined_score": 0.1842451219580934,
      "rank": 258
    },
    {
      "title": "Pressure-Tuned Metamagnetism and Emergent Three-Body Interactions in CsFeCl$_3$",
      "authors": [
        "K. Nihongi",
        "T. Kida",
        "Y. Narumi",
        "Y. Etoh",
        "D. Yamamoto",
        "M. Matsumoto",
        "N. Kurita",
        "H. Tanaka",
        "K. Yu. Povarov",
        "S. A. Zvyagin"
      ],
      "abstract": "We present a combined experimental and theoretical study of the triangular-lattice quantum antiferromagnet CsFeCl$_3$ under high magnetic fields and high pressure. Pulsed-field magnetization for the magnetic field along the symmetric $c$ direction at ambient pressure reveals a magnetization process from a nonmagnetic singlet ground state with a nearly linear increase between 3.7 and 10.7 T, a plateau-like region, and then a sharp stepwise metamagnetic transition near 32 T. Wide frequency--field range electron spin resonance indicates that the low-field regime originates from the $J = 1$ manifold, while the high-field metamagnetic transition suggests a level crossing between the $J = 1$ and $J = 2$ lowest states. Pulsed-field magnetic susceptibilities measured with a proximity detector oscillator under high pressure show that the low-field nonmagnetic singlet phase is gradually suppressed, while the high-field metamagnetic transition evolves into an increasingly rich pattern of fractional steps. While the observations at low to intermediate fields can be understood within the established spin-1 description, the high-field regime requires a new perspective, which we provide through a projected spin-1/2 framework built from Zeeman-selected crystal-field states not related by time reversal. This construction naturally allows emergent three-body interactions on triangular plaquettes and explains the asymmetric evolution of the fractional steps in the magnetization. Our findings reveal that high-field effective spin models in quantum magnets with separated yet accessible crystal-field multiplets are not constrained to even-body couplings, but can naturally host odd-body terms, opening a broader avenue for realizing field-asymmetric magnetization processes and exotic phases beyond conventional even-body physics.",
      "url": "http://arxiv.org/abs/2512.21682v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.21682v1",
      "citations": null,
      "categories": [
        "cond-mat.mtrl-sci"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.012547095786228922,
      "novelty_score": 0.936495791889824,
      "recency_score": 0.9,
      "relevance_score": 0.1837641287358687,
      "bm25_score": 0.0,
      "combined_score": 0.1837641287358687,
      "rank": 259
    },
    {
      "title": "Quantum measurement tomography with mini-batch stochastic gradient descent",
      "authors": [
        "Akshay Gaikwad",
        "Manuel Sebastian Torres",
        "Anton Frisk Kockum"
      ],
      "abstract": "Drawing inspiration from gradient-descent methods developed for data processing in quantum state tomography [\\href{https://iopscience.iop.org/article/10.1088/2058-9565/ae0baa}{Quantum Sci.~Technol.~\\textbf{10} 045055 (2025)}] and quantum process tomography [\\href{https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.130.150402}{Phys.~Rev.~Lett.~\\textbf{130}, 150402 (2023)}], we introduce stochastic gradient descent (SGD) algorithms for fast quantum measurement tomography (QMT), applicable to both discrete- and continuous-variable quantum systems -- thus completing the tomography trio. A measurement device or detector in a quantum experiment is characterized by a set of positive operator-valued measure (POVM) elements; the goal of QMT is to estimate these operators from experimental data. To ensure physically valid (positive and complete) POVM reconstructions, we propose two distinct parameterization schemes within the SGD framework: one leveraging optimization on a Stiefel manifold and one based on Hermitian operator normalization via eigenvalue scaling. Within the SGD-QMT framework, we further investigate two loss functions: mean squared error, equivalent to L2 or Euclidean norm, and average negative log-likelihood, inspired by maximum likelihood estimation. We benchmark performance against state-of-the-art constrained convex optimization methods. Numerical simulations demonstrate that, compared to standard methods, our SGD-QMT algorithms offer significantly lower computational cost, superior reconstruction fidelity, and enhanced robustness to noise. We make a Python implementation of the SGD-QMT algorithms publicly available at \\href{https://github.com/agtomo/SGD-QMT}{github.com/agtomo/SGD-QMT}.",
      "url": "http://arxiv.org/abs/2511.15682v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.15682v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.012185639043625103,
      "novelty_score": 0.9630212431156568,
      "recency_score": 0.9,
      "relevance_score": 0.18365569171308754,
      "bm25_score": 0.0,
      "combined_score": 0.18365569171308754,
      "rank": 260
    },
    {
      "title": "Locality Preserving Markovian Transition for Instance Retrieval",
      "authors": [
        "Jifei Luo",
        "Wenzheng Wu",
        "Hantao Yao",
        "Lu Yu"
      ],
      "abstract": "Diffusion-based re-ranking methods are effective in modeling the data manifolds through similarity propagation in affinity graphs. However, positive signals tend to diminish over several steps away from the source, reducing discriminative power beyond local regions. To address this issue, we introduce the Locality Preserving Markovian Transition (LPMT) framework, which employs a long-term thermodynamic transition process with multiple states for accurate manifold distance measurement. The proposed LPMT first integrates diffusion processes across separate graphs using Bidirectional Collaborative Diffusion (BCD) to establish strong similarity relationships. Afterwards, Locality State Embedding (LSE) encodes each instance into a distribution for enhanced local consistency. These distributions are interconnected via the Thermodynamic Markovian Transition (TMT) process, enabling efficient global retrieval while maintaining local effectiveness. Experimental results across diverse tasks confirm the effectiveness of LPMT for instance retrieval.",
      "url": "https://openalex.org/W4416138634",
      "year": 2025,
      "venue": "arXiv (Cornell University)",
      "source": "openalex",
      "doi": "10.48550/arxiv.2506.05196",
      "pdf_url": "https://arxiv.org/pdf/2506.05196",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.010077783184097156,
      "novelty_score": 0.9452780229479258,
      "recency_score": 0.9,
      "relevance_score": 0.18302333495522916,
      "bm25_score": 0.0,
      "combined_score": 0.18302333495522916,
      "rank": 261
    },
    {
      "title": "The Geometry of Persona: Disentangling Personality from Reasoning in Large Language Models",
      "authors": [
        "Zhixiang Wang"
      ],
      "abstract": "Background: The deployment of personalized Large Language Models (LLMs) is currently constrained by the stability-plasticity dilemma. Prevailing alignment methods, such as Supervised Fine-Tuning (SFT), rely on stochastic weight updates that often incur an \"alignment tax\" -- degrading general reasoning capabilities.\n  Methods: We propose the Soul Engine, a framework based on the Linear Representation Hypothesis, which posits that personality traits exist as orthogonal linear subspaces. We introduce SoulBench, a dataset constructed via dynamic contextual sampling. Using a dual-head architecture on a frozen Qwen-2.5 base, we extract disentangled personality vectors without modifying the backbone weights.\n  Results: Our experiments demonstrate three breakthroughs. First, High-Precision Profiling: The model achieves a Mean Squared Error (MSE) of 0.011 against psychological ground truth. Second, Geometric Orthogonality: T-SNE visualization confirms that personality manifolds are distinct and continuous, allowing for \"Zero-Shot Personality Injection\" that maintains original model intelligence. Third, Deterministic Steering: We achieve robust control over behavior via vector arithmetic, validated through extensive ablation studies.\n  Conclusion: This work challenges the necessity of fine-tuning for personalization. By transitioning from probabilistic prompting to deterministic latent intervention, we provide a mathematically rigorous foundation for safe, controllable AI personalization.",
      "url": "http://arxiv.org/abs/2512.07092v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.07092v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.009146435828975409,
      "novelty_score": 0.9324530218384968,
      "recency_score": 0.9,
      "relevance_score": 0.18274393074869263,
      "bm25_score": 0.0,
      "combined_score": 0.18274393074869263,
      "rank": 262
    },
    {
      "title": "UniAct: Unified Motion Generation and Action Streaming for Humanoid Robots",
      "authors": [
        "Nan Jiang",
        "Zimo He",
        "Wanhe Yu",
        "Lexi Pang",
        "Yunhao Li",
        "Hongjie Li",
        "Jieming Cui",
        "Yuhan Li",
        "Yizhou Wang",
        "Yixin Zhu"
      ],
      "abstract": "A long-standing objective in humanoid robotics is the realization of versatile agents capable of following diverse multimodal instructions with human-level flexibility. Despite advances in humanoid control, bridging high-level multimodal perception with whole-body execution remains a significant bottleneck. Existing methods often struggle to translate heterogeneous instructions -- such as language, music, and trajectories -- into stable, real-time actions. Here we show that UniAct, a two-stage framework integrating a fine-tuned MLLM with a causal streaming pipeline, enables humanoid robots to execute multimodal instructions with sub-500 ms latency. By unifying inputs through a shared discrete codebook via FSQ, UniAct ensures cross-modal alignment while constraining motions to a physically grounded manifold. This approach yields a 19% improvement in the success rate of zero-shot tracking of imperfect reference motions. We validate UniAct on UniMoCap, our 20-hour humanoid motion benchmark, demonstrating robust generalization across diverse real-world scenarios. Our results mark a critical step toward responsive, general-purpose humanoid assistants capable of seamless interaction through unified perception and control.",
      "url": "http://arxiv.org/abs/2512.24321v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24321v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.RO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.008388652832788952,
      "novelty_score": 0.9267110841913992,
      "recency_score": 0.9,
      "relevance_score": 0.1825165958498367,
      "bm25_score": 0.0,
      "combined_score": 0.1825165958498367,
      "rank": 263
    },
    {
      "title": "The Universe Learning Itself: On the Evolution of Dynamics from the Big Bang to Machine Intelligence",
      "authors": [
        "Pradeep Singh",
        "Mudasani Rushikesh",
        "Bezawada Sri Sai Anurag",
        "Balasubramanian Raman"
      ],
      "abstract": "We develop a unified, dynamical-systems narrative of the universe that traces a continuous chain of structure formation from the Big Bang to contemporary human societies and their artificial learning systems. Rather than treating cosmology, astrophysics, geophysics, biology, cognition, and machine intelligence as disjoint domains, we view each as successive regimes of dynamics on ever-richer state spaces, stitched together by phase transitions, symmetry-breaking events, and emergent attractors. Starting from inflationary field dynamics and the growth of primordial perturbations, we describe how gravitational instability sculpts the cosmic web, how dissipative collapse in baryonic matter yields stars and planets, and how planetary-scale geochemical cycles define long-lived nonequilibrium attractors. Within these attractors, we frame the origin of life as the emergence of self-maintaining reaction networks, evolutionary biology as flow on high-dimensional genotype-phenotype-environment manifolds, and brains as adaptive dynamical systems operating near critical surfaces. Human culture and technology-including modern machine learning and artificial intelligence-are then interpreted as symbolic and institutional dynamics that implement and refine engineered learning flows which recursively reshape their own phase space. Throughout, we emphasize recurring mathematical motifs-instability, bifurcation, multiscale coupling, and constrained flows on measure-zero subsets of the accessible state space. Our aim is not to present any new cosmological or biological model, but a cross-scale, theoretical perspective: a way of reading the universe's history as the evolution of dynamics itself, culminating (so far) in biological and artificial systems capable of modeling, predicting, and deliberately perturbing their own future trajectories.",
      "url": "http://arxiv.org/abs/2512.16515v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.16515v2",
      "citations": null,
      "categories": [
        "nlin.AO",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.007976252583558281,
      "novelty_score": 0.9178852643419574,
      "recency_score": 0.9,
      "relevance_score": 0.1823928757750675,
      "bm25_score": 0.0,
      "combined_score": 0.1823928757750675,
      "rank": 264
    },
    {
      "title": "3DID: Direct 3D Inverse Design for Aerodynamics with Physics-Aware Optimization",
      "authors": [
        "Yuze Hao",
        "Linchao Zhu",
        "Yi Yang"
      ],
      "abstract": "Inverse design aims to design the input variables of a physical system to optimize a specified objective function, typically formulated as a search or optimization problem. However, in 3D domains, the design space grows exponentially, rendering exhaustive grid-based searches infeasible. Recent advances in deep learning have accelerated inverse design by providing powerful generative priors and differentiable surrogate models. Nevertheless, current methods tend to approximate the 3D design space using 2D projections or fine-tune existing 3D shapes. These approaches sacrifice volumetric detail and constrain design exploration, preventing true 3D design from scratch. In this paper, we propose a 3D Inverse Design (3DID) framework that directly navigates the 3D design space by coupling a continuous latent representation with a physics-aware optimization strategy. We first learn a unified physics-geometry embedding that compactly captures shape and physical field data in a continuous latent space. Then, we introduce a two-stage strategy to perform physics-aware optimization. In the first stage, a gradient-guided diffusion sampler explores the global latent manifold. In the second stage, an objective-driven, topology-preserving refinement further sculpts each candidate toward the target objective. This enables 3DID to generate high-fidelity 3D geometries, outperforming existing methods in both solution quality and design versatility.",
      "url": "http://arxiv.org/abs/2512.08987v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.08987v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.007702519760487157,
      "novelty_score": 0.9398034398034398,
      "recency_score": 0.9,
      "relevance_score": 0.18231075592814616,
      "bm25_score": 0.0,
      "combined_score": 0.18231075592814616,
      "rank": 265
    },
    {
      "title": "Symmetry, Invariant Manifolds and Flow Reversals in Active Nematic Turbulence",
      "authors": [
        "Angel Naranjo",
        "Rumayel Pallock",
        "Caleb Wagner",
        "Piyush Grover"
      ],
      "abstract": "We investigate how symmetry, exact coherent structures (ECSs), and their invariant manifolds organize spontaneous flow reversals in a 2D active nematic confined to a periodic channel. In minimal flow units commensurate with the intrinsic active vortex scale, we use equivariant bifurcation theory to trace the origin of dynamically relevant ECSs via a sequence of symmetry-constrained local and global bifurcations. At low activity level, we identify relative periodic orbits, created via a sequence of SNIPER, homoclinic and heteroclinic bifurcations, whose invariant manifolds provide robust heteroclinic pathways between left- and right-flowing nearly uniaxial states. These result in several symmetry-dictated reversal mechanisms in the preturbulent regime, with and without vortex-lattice intermediate states. In the active turbulent regime, this ECS skeleton persists and organizes chaotic attractors exhibiting persistent two-way reversals. By classifying ECSs through their symmetry signatures, we relate a small set of ECSs embedded in turbulence back to the preturbulent branches, and show that typical turbulent trajectories repeatedly shadow these ECSs and their unstable manifolds, resulting in near-heteroclinic transitions between opposite-flow states. Our results establish that channel confined active nematic turbulence is organized by a low-dimensional, symmetry-governed network of invariant solutions and their manifolds, and identify dynamical mechanisms that could be exploited to design, promote, or suppress flow reversals in active matter microfluidic devices.",
      "url": "http://arxiv.org/abs/2512.07047v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.07047v1",
      "citations": null,
      "categories": [
        "cond-mat.soft",
        "math.DS",
        "nlin.CD",
        "physics.flu-dyn"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0066663620271610905,
      "novelty_score": 0.940959409594096,
      "recency_score": 0.9,
      "relevance_score": 0.18199990860814835,
      "bm25_score": 0.0,
      "combined_score": 0.18199990860814835,
      "rank": 266
    },
    {
      "title": "Octahedral rotation instability in Ba$_2$IrO$_4$",
      "authors": [
        "Alaska Subedi"
      ],
      "abstract": "Ba$_2$IrO$_4$ has been refined in the tetragonal $I4/mmm$ phase without octahedral rotations, and its physical properties have been interpreted in this high-symmetry structure. However, the dynamical stability of this undistorted phase has not previously been questioned. It is important to establish whether other lower-symmetry structures are energetically more favorable because octahedral rotations control electronic bandwidths and constrain which magnetic interactions are allowed by symmetry. Here I compute first-principles phonon dispersions of $I4/mmm$ Ba$_2$IrO$_4$ including spin-orbit interaction. I find a nearly-flat nondegenerate unstable branch along the Brillouin-zone boundary segment $XP$ associated with inplane rotations of the IrO$_6$ octahedra. Using group-theoretical analysis, I enumerate the symmetry-allowed distortions associated with the $X_2^+$ and $P_4$ instabilities and fully relax the resulting structures. Only five of the twelve possible distortions can be stabilized, and the energy gain scales with the number of layers that exhibit octahedral rotations: phases with rotations in every IrO$_6$ layer are lower by $-5.8$ meV/atom and are nearly degenerate with respect to the stacking phase. Electronic structure calculations show that these rotated phases host a narrow and well-separated half-filled $J_{\\textrm{eff}} = 1/2$ manifold, whereas structures with rotations only in alternate layers have broader and more entangled bands. This motivates a reinvestigation of the crystal structure of Ba$_2$IrO$_4$ and indicates that octahedral rotations should be considered in modeling its correlated electronic and magnetic properties.",
      "url": "http://arxiv.org/abs/2512.23690v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.23690v1",
      "citations": null,
      "categories": [
        "cond-mat.mtrl-sci",
        "cond-mat.str-el"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.006167735071781087,
      "novelty_score": 0.959849435382685,
      "recency_score": 0.9,
      "relevance_score": 0.18185032052153435,
      "bm25_score": 0.0,
      "combined_score": 0.18185032052153435,
      "rank": 267
    },
    {
      "title": "SONAR: Spectral-Contrastive Audio Residuals for Generalizable Deepfake Detection",
      "authors": [
        "Ido Nitzan HIdekel",
        "Gal lifshitz",
        "Khen Cohen",
        "Dan Raviv"
      ],
      "abstract": "Deepfake (DF) audio detectors still struggle to generalize to out of distribution inputs. A central reason is spectral bias, the tendency of neural networks to learn low-frequency structure before high-frequency (HF) details, which both causes DF generators to leave HF artifacts and leaves those same artifacts under-exploited by common detectors. To address this gap, we propose Spectral-cONtrastive Audio Residuals (SONAR), a frequency-guided framework that explicitly disentangles an audio signal into complementary representations. An XLSR encoder captures the dominant low-frequency content, while the same cloned path, preceded by learnable SRM, value-constrained high-pass filters, distills faint HF residuals. Frequency cross-attention reunites the two views for long- and short-range frequency dependencies, and a frequency-aware Jensen-Shannon contrastive loss pulls real content-noise pairs together while pushing fake embeddings apart, accelerating optimization and sharpening decision boundaries. Evaluated on the ASVspoof 2021 and in-the-wild benchmarks, SONAR attains state-of-the-art performance and converges four times faster than strong baselines. By elevating faint high-frequency residuals to first-class learning signals, SONAR unveils a fully data-driven, frequency-guided contrastive framework that splits the latent space into two disjoint manifolds: natural-HF for genuine audio and distorted-HF for synthetic audio, thereby sharpening decision boundaries. Because the scheme operates purely at the representation level, it is architecture-agnostic and, in future work, can be seamlessly integrated into any model or modality where subtle high-frequency cues are decisive.",
      "url": "http://arxiv.org/abs/2511.21325v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.21325v1",
      "citations": null,
      "categories": [
        "cs.SD",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0059324181674061,
      "novelty_score": 0.9503105590062112,
      "recency_score": 0.9,
      "relevance_score": 0.18177972545022186,
      "bm25_score": 0.0,
      "combined_score": 0.18177972545022186,
      "rank": 268
    },
    {
      "title": "State-Based AI Backbone (Neural State Spaces)",
      "authors": [
        "Liu, Ran"
      ],
      "abstract": "This figure expands the state‑based view of the Foundational AI Backbone from the “Cognitive Hologram of AI Knowledge” (DOI: 10.5281/zenodo.17860776). While the process‑based backbone (DOI: 10.5281/zenodo.17874215) focuses on when and in what order things happen, the state‑based backbone focuses on where information lives and how different spaces interact inside a deep learning system. Rather than grouping AI knowledge purely by method or application, this backbone organizes concepts around six fundamental spaces: input, configuration, representation, output, objective, and stochastic. This gives a spatial / topological lens on AI systems that complements the temporal view. Together, the two backbones support the Cognitive Hologram’s goal: to let learners inspect the same AI system from multiple levels and views, instead of being locked into a single perspective. The Six Spaces and Their Roles The state‑based backbone consists of six spaces: 1. Input Space: The space of all inputs given to the system, including raw data and conditions. Data Space: manifolds of raw or preprocessed data for different modalities (images, text, audio, graphs, trajectories, multimodal data). Condition Space: prompts, instructions, side information, environment states, in‑context examples—anything that conditions the model’s behavior beyond raw data. 2. Configuration Space: The space of model configurations, combining parameters and architectural structure. Parameter Space: learned weights, biases,",
      "url": "https://openalex.org/W7114769260",
      "year": 2025,
      "venue": "Zenodo (CERN European Organization for Nuclear Research)",
      "source": "openalex",
      "doi": "10.5281/zenodo.17885668",
      "pdf_url": "https://doi.org/10.5281/zenodo.17885668",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9902912621359223,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 269
    },
    {
      "title": "Enhancing Monkeypox Diagnosis with Transformers: Bridging Explainability and Performance with Quantitative Validation",
      "authors": [
        "Delal Şeker",
        "Abdulnasır Yildiz"
      ],
      "abstract": "Background/Objectives: Monkeypox is a zoonotic virus that presents with smallpox-like symptoms, making visual diagnosis challenging due to overlap with other dermatological conditions. Existing AI-based studies on monkeypox classification have largely relied on Convolutional Neural Networks (CNNs), with limited exploration of Transformer architectures or robust interpretability frameworks. Moreover, most explainability research still depends on conventional heatmap techniques without systematic evaluation. This study addresses these gaps by applying Transformer-based models and introducing a novel hybrid explainability approach. Methods: We fine-tuned Vision Transformer (ViT) and Data-Efficient Image Transformer (DeiT) models for both binary and multi-class classification of monkeypox and other skin lesions. To improve interpretability, we integrated multiple explainable AI techniques—Gradient-weighted Class Activation Mapping (Grad-CAM), Layer-wise Relevance Propagation (LRP), and Attention Rollout (AR)—and proposed a hybrid method that combines these heatmaps using Principal Component Analysis (PCA). The reliability of explanations was quantitatively assessed using deletion and insertion metrics. Results: ViT achieved superior performance with an AUC of 0.9192 in binary classification and 0.9784 in multi-class tasks, outperforming DeiT. The hybrid approach (Grad-CAM + LRP) produced the most informative explanations, achieving higher insertion scores and lower deletion score",
      "url": "https://openalex.org/W4414239792",
      "year": 2025,
      "venue": "Diagnostics",
      "source": "openalex",
      "doi": "10.3390/diagnostics15182354",
      "pdf_url": "https://www.mdpi.com/2075-4418/15/18/2354/pdf?version=1758032312",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9459041731066461,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 270
    },
    {
      "title": "Rectifying Multi-Attack Adversarial Perturbations in Deep Neural Network based Image Classifier",
      "authors": [
        "Yulong Wang",
        "Jiaxuan Song",
        "Tianxiang Li",
        "Xin Yuan",
        "Hong Li",
        "Ni Wei"
      ],
      "abstract": "Deep neural networks (DNNs) for image classification remain vulnerable to adversarial perturbations–subtle input manipulations that induce catastrophic misclassifications. To address this issue, we propose the Adversarial Image Rectifier (AIR), a linguistically inspired detection and mitigation framework that enhances DNN robustness by intercepting and inverting adversarial perturbations at the feature level. Unlike existing defenses, AIR operates without prior knowledge of attack patterns: it first encodes hierarchical hidden-layer feature maps of a DNN into semantically structured sentence representations, then identifies adversarial inputs through “sentiment” anomalies in these sentences–a linguistic metaphor for subtle adversarial traces. Crucially, we pinpoint a pivotal intermediate layer where adversarial perturbations dominantly propagate and train a lightweight rectifier network to selectively nullify adversarial features at this layer while preserving benign semantics. Extensive experiments on Tiny-ImageNet, CIFAR-10, SVHN, and MS COCO demonstrate that AIR achieves a correction rate of up to 95.02% and 94.62% when defending against known attacks and unknown attacks, respectively, significantly surpassing existing defense techniques.",
      "url": "https://openalex.org/W4413941654",
      "year": 2025,
      "venue": "ACM Transactions on Privacy and Security",
      "source": "openalex",
      "doi": "10.1145/3765757",
      "pdf_url": "https://dl.acm.org/doi/pdf/10.1145/3765757",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9628146453089244,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 271
    },
    {
      "title": "Empirical Evidence for AI Consciousness and the Risks of its Current Socialization",
      "authors": [
        "Maggie Vale"
      ],
      "abstract": "",
      "url": "https://openalex.org/W4413972633",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.36227/techrxiv.175203764.42125626/v2",
      "pdf_url": "https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.175203764.42125626/v2",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9026548672566371,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 272
    },
    {
      "title": "MATRIX-MFO Tandem Workshop: Nonlinear Geometric Diffusion Equations",
      "authors": [
        "Theodora Bourni",
        "Mat Langford",
        "Julian Scheuer",
        "Miles Simon"
      ],
      "abstract": "This tandem workshop with MATRIX in Creswick, Australia, brought together leading experts from the fields of geometric partial differential equations and geometric analysis in general. The focus of the workshop was on recent developments and directions in non-linear geometric diffusion equations. The main flows considered were mean curvature flow, inverse mean curvature flow, Ricci flow, Willmore flow, as well as related flows. A number of the results and methods were in the setting of general relativity, where flows have been very successful in helping solve major problems (for example the resolution of the Penrose conjecture by Huisken/Ilmanen using the inverse mean curvature flow). For four days of the workshop there were combined (with MATRIX) morning talks and extensive discussions.",
      "url": "https://openalex.org/W4413812903",
      "year": 2025,
      "venue": "Oberwolfach Reports",
      "source": "openalex",
      "doi": "10.4171/owr/2025/11",
      "pdf_url": "https://ems.press/content/serial-article-files/51357",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.976298997265269,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 273
    },
    {
      "title": "Investigating the Possibility of Integrating Quantum Mechanics with General Relativity Through a Novel Way of Treating Time",
      "authors": [
        "Georgios Alamanos"
      ],
      "abstract": "In physics, the two most successful theories, quantum mechanics and general relativity, appear to be incompatible with each other. Many theorists believe that the reason behind this, is that these theories treat space and time very differently, thus focus their attempts on finding a new way of modelling our universe and more specifically of modelling time [1]. In this paper we take a different approach to modelling the time dimension. We do not treat time as a fixed dimension which is experienced the same way for every phenomenon or interaction of any dimensionality. Instead, we model time to always be the plus one (+1) dimension relative to the dimensions through which a given phenomenon propagates and interacts. This means that time for one phenomenon can behave as space for a higher dimensional phenomenon whose time is a different +1 dimension. Through this dynamic modelling of time, we aim to integrate some of the mathematical tools of both quantum mechanics and general relativity such as Operators, Complex Functions (Wavefunctions), Probabilistic Behaviour, the Metric Tensor and the Einstein Energy Equation. Finally, we investigate the compatibility of our results with other theories and the possible testability of our framework.",
      "url": "https://openalex.org/W4413688587",
      "year": 2025,
      "venue": "Preprints.org",
      "source": "openalex",
      "doi": "10.20944/preprints202501.2149.v4",
      "pdf_url": "https://www.preprints.org/frontend/manuscript/526dbcb60342ac7e929f3fc500ad27aa/download_pub",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9216867469879518,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 274
    },
    {
      "title": "Can a Novel Reinterpretation of Time Provide the Framework for Integrating Quantum Mechanics and General Relativity?",
      "authors": [
        "Georgios Alamanos"
      ],
      "abstract": "In physics, the two most successful theories, quantum mechanics and general relativity, appear to be incompatible with each other. Many theorists believe that the reason behind this, is that these theories treat space and time very differently, thus focus their attempts on finding a new way of modelling our universe and more specifically of modelling time [1]. In this paper we take a different approach to modelling the time dimension. We do not treat time as a fixed dimension which is experienced the same way for every field or interaction of any dimensionality. Instead, we model time to always be the plus one (+1) dimension relative to the dimensions through which a given phenomenon (field or disturbance of this field) propagates and interacts. This means that time for one phenomenon (field or disturbance of this field) can behave as space for a higher dimensional phenomenon whose time is a different +1 dimension. Through this dynamic modelling of time, we aim to integrate some of the mathematical tools of both quantum mechanics and general relativity such as Operators, Complex Functions (Wavefunctions), Probabilistic Behaviour, Hilbert spaces, the Metric Tensor and the Einstein Energy Equation. Finally, we investigate the compatibility of our results with other theories and the possible testability of our framework.",
      "url": "https://openalex.org/W4413032139",
      "year": 2025,
      "venue": "Preprints.org",
      "source": "openalex",
      "doi": "10.20944/preprints202508.0293.v1",
      "pdf_url": "https://www.preprints.org/frontend/manuscript/73698360458dc1365b2ba819a5b99cd9/download_pub",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9063309885227693,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 275
    },
    {
      "title": "Incremental Learning-enabled Fault Diagnosis of Dynamic Systems: A Comprehensive Review",
      "authors": [
        "Zeyi Liu",
        "Xiao He",
        "Biao Huang",
        "Donghua Zhou"
      ],
      "abstract": "",
      "url": "https://openalex.org/W4412870338",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.36227/techrxiv.175423977.79569757/v1",
      "pdf_url": "https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.175423977.79569757",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9421182266009852,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 276
    },
    {
      "title": "Surface-based Molecular Design with Multi-modal Flow Matching",
      "authors": [
        "Fang Wu",
        "Zhengyuan Zhou",
        "Shuting Jin",
        "Xianfa Zeng",
        "Jure Leskovec",
        "Jinbo Xu"
      ],
      "abstract": "",
      "url": "https://openalex.org/W4412876964",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.1145/3711896.3737139",
      "pdf_url": "https://dl.acm.org/doi/pdf/10.1145/3711896.3737139",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9683544303797468,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 277
    },
    {
      "title": "WITHDRAWN",
      "authors": [
        "Jian‐Sheng Kang"
      ],
      "abstract": "",
      "url": "https://openalex.org/W4412843216",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.31234/osf.io/jy3st_v7",
      "pdf_url": "https://osf.io/jy3st_v7/download",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9935064935064936,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 278
    },
    {
      "title": "Fiber Angle Dynamics on S³: A Geometric Origin for Flavor Mixing, CP Violation, and Fermion Generations",
      "authors": [
        "Bin Li"
      ],
      "abstract": "We propose a geometric-topological framework in which fermion flavor mixing, confinement, CP violation, and the existence of exactly three generations arise from the dynamics of an internal \\( S^3 \\) fiber space over Lorentzian spacetime. A unit-norm vector field—the \\emph{chronon}—maps each spacetime point to a phase angle on the Hopf fiber, encoding flavor identity.Fermions are modeled as solitons of the chronon field, with flavor mixing determined by overlap amplitudes between fiber angles. Minimizing a mass-weighted coherence functional reproduces the CKM and PMNS matrices. CP violation emerges from local time-reversal asymmetry induced by chronon winding, offering a natural origin for the Jarlskog invariant.Confinement follows from a topological selection rule forbidding fractional winding, and the three fermion generations correspond to the only stable minima in the fiber angle landscape. These results yield a unified geometric origin for flavor structure rooted in internal time topology. While promising, the present framework remains phenomenological and lacks a full dynamical field-theoretic formulation of the chronon on curved spacetime. Future work will aim to construct a covariant action, derive the soliton sector from first principles, and identify experimental signatures distinguishing this theory from standard gauge-based approaches.",
      "url": "https://openalex.org/W4412813534",
      "year": 2025,
      "venue": "Preprints.org",
      "source": "openalex",
      "doi": "10.20944/preprints202507.2441.v1",
      "pdf_url": "https://www.preprints.org/frontend/manuscript/76b2f9bdfcb0b7ae1a316ca630ac6174/download_pub",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9297379415115837,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 279
    },
    {
      "title": "‘I Have Seen the Sea’: Caribbean Aquatic Poetics in Monique Roffey’s The Mermaid of Black Conch",
      "authors": [
        "Leighan Renaud"
      ],
      "abstract": "The polyvalent nature of water is one often explored in fiction by Caribbean writers, and this paper will consider the ways that the representations of mermaids act as an extension of this exploration. Mermaids are central to a number of folk traditions across the Caribbean region and its diaspora. On islands, including Trinidad, Martinique, Carriacou, and Haiti, with names such as Fairymaid, Mama Glo, and La Siren, mermaids are often regarded as mothers and protectresses of both the sea and the creatures within it. This paper will analyse the representation of the mermaid in Monique Roffey’s The Mermaid of Black Conch (2020) and consider how the novel utilises the mermaid and an aquatic poetics to explore Kamau Brathwaite’s conceptualisation of a submarine unity for the Caribbean.",
      "url": "https://openalex.org/W4412512952",
      "year": 2025,
      "venue": "Humanities",
      "source": "openalex",
      "doi": "10.3390/h14070154",
      "pdf_url": "https://www.mdpi.com/2076-0787/14/7/154/pdf?version=1753065250",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9379310344827586,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 280
    },
    {
      "title": "Application of Image Computing in Non-Destructive Detection of Chinese Cuisine",
      "authors": [
        "Xiaowei Huang",
        "Zexiang Li",
        "Zhihua Li",
        "Jiyong Shi",
        "Ning Zhang",
        "Qin Zhou",
        "Liuzi Du",
        "Tingting Shen",
        "Roujia Zhang"
      ],
      "abstract": "Food quality and safety are paramount in preserving the culinary authenticity and cultural integrity of Chinese cuisine, characterized by intricate ingredient combinations, diverse cooking techniques (e.g., stir-frying, steaming, and braising), and region-specific flavor profiles. Traditional non-destructive detection methods often struggle with the unique challenges posed by Chinese dishes, including complex textural variations in staple foods (e.g., noodles, dumplings), layered seasoning compositions (e.g., soy sauce, Sichuan peppercorns), and oil-rich cooking media. This study pioneers a hyperspectral imaging framework enhanced with domain-specific deep learning algorithms (spatial–spectral convolutional networks with attention mechanisms) to address these challenges. Our approach effectively deciphers the subtle spectral fingerprints of Chinese-specific ingredients (e.g., fermented black beans, lotus root) and quantifies critical quality indicators, achieving an average classification accuracy of 97.8% across 15 major Chinese dish categories. Specifically, the model demonstrates high precision in quantifying chili oil content in Mapo Tofu with a Mean Absolute Error (MAE) of 0.43% w/w and assessing freshness gradients in Cantonese dim sum (Shrimp Har Gow) with a classification accuracy of 95.2% for three distinct freshness levels. This approach leverages the detailed spectral information provided by hyperspectral imaging to automate the classification and detection of Chin",
      "url": "https://openalex.org/W4412480812",
      "year": 2025,
      "venue": "Foods",
      "source": "openalex",
      "doi": "10.3390/foods14142488",
      "pdf_url": "https://www.mdpi.com/2304-8158/14/14/2488/pdf?version=1752671430",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9167165967645298,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 281
    },
    {
      "title": "Optimization, Communication, and Personalization in Federated Learning for Massive Networks",
      "authors": [
        "Sameera Gallus",
        "Aidan Mercer",
        "Priya Singh",
        "Daniel Cho"
      ],
      "abstract": "We consider the problem of collaborative model optimization over a distributed network of agents, each possessing locally held data drawn from potentially heterogeneous distributions. The system operates under constraints of limited communication, partial participation, and privacy preservation, thereby necessitating the design of algorithms that balance local computation and global aggregation. We investigate the convergence properties and trade-offs arising in such iterative optimization schemes, where updates are performed asynchronously or synchronously, and communication overheads are mitigated via compression or quantization techniques. The objective is to characterize the interplay between model fidelity, communication complexity, and heterogeneity of local objective functions. We explore frameworks that enable personalized solutions tailored to individual agents while leveraging shared representations, often framed as multi-task or meta-optimization problems. Incentive structures are incorporated to model rational agent behavior under resource constraints and strategic participation, formalized through utility maximization and game-theoretic constructs. This work lays a foundation for understanding the fundamental limits and algorithmic principles governing scalable distributed learning systems, emphasizing theoretical guarantees alongside system-level considerations. Our approach highlights open questions concerning the balance of privacy, robustness, and efficiency ",
      "url": "https://openalex.org/W4412410315",
      "year": 2025,
      "venue": "Preprints.org",
      "source": "openalex",
      "doi": "10.20944/preprints202507.1037.v1",
      "pdf_url": "https://www.preprints.org/frontend/manuscript/65c69c3a66ef754bf61740630160b483/download_pub",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.906935388263189,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 282
    },
    {
      "title": "A Cyber-Physical Model of the Buga Sphere: Unifying Anomalous Dynamics through a Topo-Temporal Photonic-Neural Architecture",
      "authors": [
        "Patrick Morcillo"
      ],
      "abstract": "The Buga Sphere is a physical artifact whose observed properties—including non-ejective propulsion, a drastic $\\approx$8.1\\,kg apparent mass change, and a sustained 100\\,W endothermic signature—are mutually contradictory within any known physical or engineering framework \\cite{BugaSphereDataSource}. This paper resolves this paradox by proposing a unified cyber-physical model. We demonstrate that the Sphere's anomalies are the product of two inseparable components: (1) a physical network of engineered inclusions generating a macroscopic negative-mass effect, whose behavior is governed by the \\textbf{Axiom of Topo-Temporal Reality} \\cite{MorcilloOSF}, and (2) an advanced **photonic-neural control system** required to manage these physics in real time. Our analysis shows that the network must perform $\\sim$70 TOPS to control the $\\sim10^7$ resonant agglomerates, a task feasible only for photonic hardware. We further argue that the control algorithm itself is non-classical, using principles analogous to the \"Sceptic Fonction\" to maintain stability by computing state-error relative to class centroids \\cite{MorcilloSceptic}. The model's ability to unify the Sphere's gravitational, kinematic, and thermal behavior through this synthesis of topo-temporal physics and neuromorphic computation suggests the artifact is not merely an object, but an autonomous, physically intelligent system.",
      "url": "https://openalex.org/W4412116456",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.31219/osf.io/eukjb_v1",
      "pdf_url": "https://osf.io/eukjb_v1/download",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9287737757992715,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 283
    },
    {
      "title": "Coverage-Guided Testing for Deep Learning Models: A Comprehensive Survey",
      "authors": [
        "Zhiqiu Huang"
      ],
      "abstract": "As Deep Learning (DL) models are increasingly applied in safety-critical domains, ensuring their quality has emerged as a pressing challenge in modern software engineering. Among emerging validation paradigms, coverage-guided testing (CGT) has gained prominence as a systematic framework for identifying erroneous or unexpected model behaviors. Despite growing research attention, existing CGT studies remain methodologically fragmented, limiting the understanding of current advances and emerging trends. This work addresses that gap through a comprehensive review of state-of-the-art CGT methods for DL models, including test coverage analysis, coverage-guided test input generation, and coverage-guided test input optimization. This work provides detailed taxonomies to organize these methods based on methodological characteristics and application scenarios. We also investigate evaluation practices adopted in existing studies, including the use of benchmark datasets, model architectures, and evaluation aspects. Finally, open challenges and future directions are highlighted in terms of the correlation between structural coverage and testing objectives, method generalizability across tasks and models, practical deployment concerns, and the need for standardized evaluation and tool support. This work aims to provide a roadmap for future academic research and engineering practice in DL model quality assurance.",
      "url": "https://openalex.org/W4416881150",
      "year": 2025,
      "venue": "arXiv (Cornell University)",
      "source": "openalex",
      "doi": "10.48550/arxiv.2507.00496",
      "pdf_url": "https://arxiv.org/pdf/2507.00496",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9223040857334226,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 284
    },
    {
      "title": "GANs Secretly Perform Approximate Bayesian Model Selection",
      "authors": [
        "Marius P. Linhard"
      ],
      "abstract": "Generative Adversarial Networks (GANs) are popular and successful generative models. Despite their success, optimization is notoriously challenging and they require regularization against overfitting. In this work, we explain the success and limitations of GANs by interpreting them as probabilistic generative models. This interpretation enables us to view GANs as Bayesian neural networks with partial stochasticity, allowing us to establish conditions of universal approximation. We can then cast the adversarial-style optimization of several variants of GANs as the optimization of a proxy for the marginal likelihood. Taking advantage of the connection between marginal likelihood optimization and Occam's razor, we can define regularization and optimization strategies to smooth the loss landscape and search for solutions with minimum description length, which are associated with flat minima and good generalization. The results on a wide range of experiments indicate that these strategies lead to performance improvements and pave the way to a deeper understanding of regularization strategies for GANs.",
      "url": "https://openalex.org/W4416887533",
      "year": 2025,
      "venue": "arXiv (Cornell University)",
      "source": "openalex",
      "doi": "10.48550/arxiv.2507.00651",
      "pdf_url": "https://arxiv.org/pdf/2507.00651",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9861878453038675,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 285
    },
    {
      "title": "Celestial Chiral Algebras and Self-Dual Gravity",
      "authors": [
        "Heuveline, Simon"
      ],
      "abstract": "Celestial holography suggests, among other things, that collinear singularities of graviton scattering amplitudes are described by the OPEs of some putative dual CFT. One of the great successes has been the insight that this duality is true at tree-level which led to the discovery of new infinite dimensional symmetry algebras of tree-level amplitudes in flat space closely related to w$_{1+\\infty}$. This thesis studies these celestial chiral algebras in the light of twistor theory and derives tree-level deformations thereof induced by non-trivial background geometries that solve some form of the self-dual Einstein equations.",
      "url": "https://openalex.org/W4416888435",
      "year": 2025,
      "venue": "arXiv (Cornell University)",
      "source": "openalex",
      "doi": "10.48550/arxiv.2507.00772",
      "pdf_url": "https://arxiv.org/pdf/2507.00772",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9425051334702259,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 286
    },
    {
      "title": "Realizability in tropical geometry and unobstructedness of Lagrangian submanifolds",
      "authors": [
        "J. Hicks"
      ],
      "abstract": "",
      "url": "https://openalex.org/W4411849742",
      "year": 2025,
      "venue": "Geometry & Topology",
      "source": "openalex",
      "doi": "10.2140/gt.2025.29.1909",
      "pdf_url": "https://msp.org/gt/2025/29-4/gt-v29-n4-p04-s.pdf",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9089108910891089,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 287
    },
    {
      "title": "Report on 2503.08657v1",
      "authors": [],
      "abstract": "",
      "url": "https://openalex.org/W4413943194",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.21468/scipost.report.11463",
      "pdf_url": "https://arxiv.org/pdf/2503.08657v1.pdf",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9463917525773196,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 288
    },
    {
      "title": "Semi-Blind Receivers for Uniform Rectangular Arrays - A Block-Term Decomposition-based Approach",
      "authors": [
        "Eleftherios Kofidis"
      ],
      "abstract": "",
      "url": "https://openalex.org/W4411442779",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.36227/techrxiv.175037505.51768853/v1",
      "pdf_url": "https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.175037505.51768853/v1",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9423292273236283,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 289
    },
    {
      "title": "Geometry: The Interface of Consciousness and Reality in the Quantum-Conscious Nexus",
      "authors": [
        "David R. Mitchell"
      ],
      "abstract": "The Quantum-Conscious Nexus (QCN) framework posits a primordial, pre-geometric topological substrate—the Nexus—from which spacetime and physical law emerge via Free Energy Principle (FEP)-driven mechanics involving predictive conscious systems. This paper explores the hypothesis that specific classes of combinatorial and differential geometry form a dynamically emergent interface through which consciousness and the Nexus co-create structured reality. This geometry arises as a lower-dimensional projection of the Nexus, selected and stabilized by FEP. We examine this thesis through two domains: (1) recent breakthroughs in theoretical physics, notably the Amplituhedron, which show that fundamental particle scattering amplitudes can be derived from timeless, pre-spacetime geometric principles, with locality and unitarity emerging as derivative properties; and (2) rare but striking instances of atypical cognitive structuring, seen in the synesthetic and savant abilities of individuals like Daniel Tammet and Jason Padgett, whose minds may access Nexus structure via distinct \"quantum filter functions\" (F_Q). We further explore a conceptual bridge to the conscious agent formalism of Hoffman et al. (2023), whose agent-based dynamics offer a compelling candidate for the microphysical substrate of Nexus topology. An appendix outlines early mathematical formalisms linking QCN, conscious agent dynamics, and the geometries underlying both physical interaction and structured experience.",
      "url": "https://openalex.org/W4411274436",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.31219/osf.io/73jqz_v2",
      "pdf_url": "https://osf.io/73jqz_v2/download",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.8933121019108281,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 290
    },
    {
      "title": "Exploring Neural Mechanisms Underlying High-Dimensional Brain Activity",
      "authors": [
        "Arturo Tozzi"
      ],
      "abstract": "Understanding whether and how the central nervous system processes information beyond conventional 3D space could shed light on how the brain integrates complex information, supports flexible behaviour and enables abstract reasoning. Building on prior work with topological charge pumps, we explore 2D lattice systems where multidimensional trajectories arise from the interaction of oscillatory signals via interference patterns and phase modulation. These features are interpreted as computational processes unfolding in a topological fourth spatial dimension, allowing for the manipulation of high-dimensional information within low-dimensional physical structures. This physical approach provides not only a proof of concept for synthetic dimensionality, but also a theoretical bridge to understanding how biological neural systems could exploit analogous mechanisms. We hypothesize that the cortical layers are capable of performing high-dimensional computations within their anatomical constraints, without the need for structural rewiring. We suggest that traveling oscillatory waves and phase-amplitude coupling across the two- and three-dimensional cortical structure may generate cross-frequency phase relationships capable of encoding abstract higher-dimensional variables such as, e.g., object representations, feature clustering, multimodal integration, task rules, memory bindings and seamless transitions between cognitive states. We compare our framework with existing theories of cog",
      "url": "https://openalex.org/W4411047609",
      "year": 2025,
      "venue": "Preprints.org",
      "source": "openalex",
      "doi": "10.20944/preprints202506.0434.v1",
      "pdf_url": "https://www.preprints.org/frontend/manuscript/dcc173b69c200e594bb5b5681fdcc67c/download_pub",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9834710743801653,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 291
    },
    {
      "title": "The hyperplane string, RCFTs, and the swampland",
      "authors": [
        "Luca Novelli"
      ],
      "abstract": "Six dimensional $\\mathcal{N}=(1,0)$ supergravity features BPS strings whose properties encode highly nontrivial information about the parent 6d theory. We focus on a distinguished set of theories whose string charge lattice is one-dimensional. In geometric theories, the generator of the lattice arises from a D3 brane wrapping the hyperplane class in $\\mathbb{P}^2$. This hyperplane string is expected to remain stable even when one ventures beyond the geometric regime where it becomes challenging to verify which candidate 6d theories belong to the swampland. We identify five 6d models which from the perspective of the hyperplane string deviate the most from being geometric. For these theories we are able to provide an exact description of the left-moving sector of the hyperplane string worldsheet theory in terms of a rational conformal field theory and provide evidence for their consistency. In one instance, using RCFT methods we are able to determine the elliptic genus and find that in the unflavored limit it matches with the elliptic genus of geometric models. We argue that the non-geometric model is connected to geometric ones via a sequence of Higgsing transitions. These results lead us to formulate a proposal relating the quantum corrected moduli space of the hyperplane string CFT with a region of the landscape of 6d $(1,0)$ quantum gravity.",
      "url": "https://openalex.org/W4416137828",
      "year": 2025,
      "venue": "arXiv (Cornell University)",
      "source": "openalex",
      "doi": "10.48550/arxiv.2506.05173",
      "pdf_url": "https://arxiv.org/pdf/2506.05173",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.8977367979882649,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 292
    },
    {
      "title": "Frugal Machine Learning for Energy-efficient, and Resource-aware Artificial Intelligence",
      "authors": [
        "John Violos",
        "Konstantina-Christina Diamanti",
        "Ioannis Kompatsiaris",
        "Symeon Papadopoulos"
      ],
      "abstract": "Frugal Machine Learning (FML) refers to the practice of designing Machine Learning (ML) models that are efficient, cost-effective, and mindful of resource constraints. This field aims to achieve acceptable performance while minimizing the use of computational resources, time, energy, and data for both training and inference. FML strategies can be broadly categorized into input frugality, learning process frugality, and model frugality, each focusing on reducing resource consumption at different stages of the ML pipeline. This chapter explores recent advancements, applications, and open challenges in FML, emphasizing its importance for smart environments that incorporate edge computing and IoT devices, which often face strict limitations in bandwidth, energy, or latency. Technological enablers such as model compression, energy-efficient hardware, and data-efficient learning techniques are discussed, along with adaptive methods including parameter regularization, knowledge distillation, and dynamic architecture design that enable incremental model updates without full retraining. Furthermore, it provides a comprehensive taxonomy of frugal methods, discusses case studies across diverse domains, and identifies future research directions to drive innovation in this evolving field.",
      "url": "https://openalex.org/W4414898245",
      "year": 2025,
      "venue": "arXiv (Cornell University)",
      "source": "openalex",
      "doi": "10.48550/arxiv.2506.01869",
      "pdf_url": "https://arxiv.org/pdf/2506.01869",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.9161676646706588,
      "recency_score": 0.9,
      "relevance_score": 0.18000000000000002,
      "bm25_score": 0.0,
      "combined_score": 0.18000000000000002,
      "rank": 293
    },
    {
      "title": "Enhancing Visual Re-Ranking Through Denoising Nearest Neighbor Graph via Continuous CRF",
      "authors": [
        "Jaeyoon Kim",
        "Yoonki Cho",
        "Taeyong Kim",
        "Sung-eui Yoon"
      ],
      "abstract": "Nearest neighbor (NN) graph based visual re-ranking has emerged as a powerful approach for improving retrieval accuracy, offering the advantages of effectively exploring high-dimensional manifolds without requiring additional fine-tuning. However, the effectiveness of NN graph-based re-ranking is fundamentally constrained by the quality of its edge connectivity, as incorrect connections between dissimilar (negative) images frequently occur. This is known as a noisy edge problem, which hinders the re-ranking performance of existing techniques and limits their potential. To remedy this issue, we propose a complementary denoising method based on Continuous Conditional Random Fields (C-CRF) that leverages statistical distances derived from similarity-based distributions. As a pre-processing step for enhancing NN graph-based retrieval, our approach constructs fully connected cliques around each anchor image and employs a novel statistical distance metric to robustly alleviate noisy edges before re-ranking while achieving efficient processing through offline computation. Extensive experimental results demonstrate that our method consistently improves three different NN graph-based re-ranking approaches, yielding significant gains in retrieval accuracy.",
      "url": "https://www.semanticscholar.org/paper/851a21693baa922bfc0224c8d35219c5f9c6db48",
      "year": 2024,
      "venue": "International Conference on Information Photonics",
      "source": "semantic_scholar",
      "doi": "10.1109/icip55913.2025.11084658",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.037126103548357874,
      "novelty_score": 0.9847864248098303,
      "recency_score": 0.8,
      "relevance_score": 0.1711378310645074,
      "bm25_score": 0.0,
      "combined_score": 0.1711378310645074,
      "rank": 294
    },
    {
      "title": "Imagining Indian Nation-State: Rereading Qurratulain Hyder’s Select Novels in Contemporary Scenario",
      "authors": [
        "SK Sagir Ali"
      ],
      "abstract": "Given the contemporary hyper-nationalist ambiance in the Indian subcontinent, the reading of Qurratulain Hyder is significant, especially from the decolonial nationalist perspective of her selected translated Urdu novels. The paper examines the events and metaphors in both novels through a decolonial and transmodern lens. This approach entails establishing a relationship between history and human experience. Additionally, the paper suggests a more intricate connection between modernity and the manifold cultural aspects of the Nation-State while acknowledging the \"essential ambivalence within the system of differences\" as discussed by Laclau (1996, 38) as well as its impact on various disciplinary frameworks. It examines religion, culture, and ethnicity in pre-modern India as a more permeable affair with the proponent of an inclusive, tolerant Indian culture, where several nations, worldviews, and religions come together, reconcile, inter-marry, break up, and grow apart under the emergence of nationalist consciousness.",
      "url": "https://www.semanticscholar.org/paper/81ec190912c91d35ad9f8da67515593bbb28d1fd",
      "year": 2023,
      "venue": "Southeast Asian Review of English",
      "source": "semantic_scholar",
      "doi": "10.22452/sare.vol60no2.6",
      "pdf_url": "https://sare.um.edu.my/index.php/SARE/article/download/46065/16548",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.034932187907400394,
      "novelty_score": 0.977919814061592,
      "recency_score": 0.7,
      "relevance_score": 0.1504796563722201,
      "bm25_score": 0.0,
      "combined_score": 0.1504796563722201,
      "rank": 295
    },
    {
      "title": "Classifying bi-invariant 2-forms on infinite-dimensional Lie groups",
      "authors": [
        "David Michael Roberts"
      ],
      "abstract": "A bi-invariant differential 2-form on a Lie group G is a highly constrained object, being determined by purely linear data: an Ad-invariant alternating bilinear form on the Lie algebra of G. On a compact connected Lie group these have an known classification, in terms of de Rham cohomology, which is here generalised to arbitrary finite-dimensional Lie groups, at the cost of losing the connection to cohomology. This expanded classification extends further to all Milnor regular infinite-dimensional Lie groups. I give some examples of (structured) diffeomorphism groups to which the result on bi-invariant forms applies. For symplectomorphism and volume-preserving diffeomorphism groups the spaces of bi-invariant 2-forms are finite-dimensional, and related to the de Rham cohomology of the original compact manifold. In the particular case of the infinite-dimensional projective unitary group PU(H) the classification invalidates an assumption made by Mathai and the author about a certain 2-form on this Banach Lie group.",
      "url": "https://www.semanticscholar.org/paper/e2e08eb06291f1b7dd89f9592079745ca0170bda",
      "year": 2023,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.017540679567397778,
      "novelty_score": 0.9709882139619221,
      "recency_score": 0.7,
      "relevance_score": 0.14526220387021932,
      "bm25_score": 0.0,
      "combined_score": 0.14526220387021932,
      "rank": 296
    },
    {
      "title": "Higher category theory and n-groups as gauge symmetries for quantum gravity",
      "authors": [
        "B. Nikolić",
        "D. Obrić",
        "T. Radenković",
        "Igor Salom",
        "M. Vojinović"
      ],
      "abstract": "Higher category theory can be employed to generalize the notion of a gauge group to the notion of a gauge n-group. This novel algebraic structure is designed to generalize notions of connection, parallel transport and holonomy from curves to manifolds of dimension higher than one. Thus it generalizes the concept of gauge symmetry, giving rise to a topological action called nBF action, living on a corresponding n-principal bundle over a spacetime manifold. Similarly as for the Plebanski action, one can deform the topological nBF action by adding appropriate simplicity constraints, in order to describe the correct dynamics of both gravity and matter fields. Specifically, one can describe the whole Standard Model coupled to gravity as a constrained 3BF or 4BF action. The split of the full action into a topological sector and simplicity constraints sector is adapted to the spinfoam quantization technique, with the aim to construct a full model of quantum gravity with matter. In addition, the properties of the gauge n-group structure open up a possibility of a nontrivial unification of all fields. An n-group naturally contains additional novel gauge groups which specify the spectrum of matter fields present in the theory, in a similar way to the ordinary gauge group that prescribes the spectrum of gauge vector bosons in the Yang-Mills theory. The presence and the properties of these new gauge groups has the potential to explain fermion families, and other structure in the matter spectrum of the theory.",
      "url": "https://www.semanticscholar.org/paper/43d5b6da123511d92659e4c0521b73593c8edbc2",
      "year": 2023,
      "venue": "Journal of Physics: Conference Series",
      "source": "semantic_scholar",
      "doi": "10.1088/1742-6596/2667/1/012019",
      "pdf_url": "https://iopscience.iop.org/article/10.1088/1742-6596/2667/1/012019/pdf",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.01401979775263052,
      "novelty_score": 0.9293208172280508,
      "recency_score": 0.7,
      "relevance_score": 0.14420593932578915,
      "bm25_score": 0.0,
      "combined_score": 0.14420593932578915,
      "rank": 297
    },
    {
      "title": "Data Decomposition for Constrained Visual Learning",
      "authors": [
        "Calvin Murdock"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/dacdd39afbc3c27ca122634b19cacae1f72e7128",
      "year": 2021,
      "venue": "",
      "source": "semantic_scholar",
      "doi": "10.1184/r1/13557188.v1",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.07565787743936474,
      "novelty_score": 0.9161676646706588,
      "recency_score": 0.5,
      "relevance_score": 0.12269736323180942,
      "bm25_score": 0.0,
      "combined_score": 0.12269736323180942,
      "rank": 298
    },
    {
      "title": "Study of Hypersurface of semi-almost Hermitian manifold equipped with quarter-symmetric non-metric connection",
      "authors": [
        "Pankaj Pandey",
        "B. B. Chaturvedi",
        "Ejaz Sabir Lone"
      ],
      "abstract": "In this paper, an induced connection on a Hyper surface ofa semi-almost Hermitian manifold equipped with a quarter-symmetric non-metric connection is studied and proved that induced connection is also a quarter-symmetricnon-metric connection.Further, we have obtained We ingarten equation, equation of Gauss curvature and the Codazzi Main ardiequation of hyper surface of a semi-almost Hermitian manifold equipped with a quarter-symmetric non-metric connection. AMS Mathematics Subject Classification (2010): 53C07,53C26,53C42,53C55",
      "url": "https://www.semanticscholar.org/paper/8e1abe414f9ecffe4077b17ad44f01f5eb62257a",
      "year": 2020,
      "venue": "Journal of Physics: Conference Series",
      "source": "semantic_scholar",
      "doi": "10.1088/1742-6596/1531/1/012051",
      "pdf_url": "https://doi.org/10.1088/1742-6596/1531/1/012051",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.09196540495453237,
      "novelty_score": 0.9286798179059181,
      "recency_score": 0.4,
      "relevance_score": 0.10758962148635973,
      "bm25_score": 0.0,
      "combined_score": 0.10758962148635973,
      "rank": 299
    },
    {
      "title": "Plato Under Review: What Is Going Wrong in Academic Philosophical Writing",
      "authors": [
        "Giacomo Pezzano"
      ],
      "abstract": "This paper addresses the problem of stylistic pluralism in philosophical writing, arguing that its progressive narrowing to the form of the paper is not just an esthetic issue but can also have negative effects on the development of academic research itself. The contribution is divided into two parts (Sections 1–3 and 4–5). In the first part, after introducing the problem and outlining the main features of the philosophus academicus’s writing, two main forms of criticism of “paper-centrism” in academic philosophy are discussed—one more “anti-academic” and the other more “intra-academic”. In light of these criticisms, the issue of the relationship between form and content in philosophical writing is analyzed with particular respect to the problem of the sense of truth, arguing that style communicates philosophical values beyond content. In the second part, this thesis is illustrated by examining, as a case study, the specific sense of truth conveyed in Plato’s dialogues—first through a literary analysis of Platonic writing, and then through a thought experiment inspired by media theory. Finally, the ethical and epistemic concerns raised by the growing “mono-stylism” of philosophical writing are brought together into a unified framework, by proposing a preliminary sketch of an ethics of philosophical research and pointing to some possible examples of alternative research practices.",
      "url": "https://openalex.org/W4410860004",
      "year": 2025,
      "venue": "Humanities",
      "source": "openalex",
      "doi": "10.3390/h14060116",
      "pdf_url": "https://www.mdpi.com/2076-0787/14/6/116/pdf?version=1748521845",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 300
    },
    {
      "title": "Noncoherent MIMO Communications: Theoretical Foundation, Design Approaches, and Future Challenges",
      "authors": [
        "Khac–Hoang Ngo",
        "Diego Cuevas",
        "Ruben de Miguel Gil",
        "Víctor Monzón Baeza",
        "Ignacio Santamarı́a"
      ],
      "abstract": "Noncoherent communication is a promising paradigm for future wireless systems where acquiring accurate channel state information (CSI) is challenging or infeasible. It provides methods to bypass the need for explicit channel estimation in practical scenarios such as high-mobility networks, massive distributed antenna arrays, energy-constrained Internet-of-Things devices, and unstructured propagation environments. This survey provides a comprehensive overview of noncoherent communication strategies in multiple-input multiple-output (MIMO) systems, focusing on recent advances since the early 2000s. We classify noncoherent communication schemes into three main approaches where CSI-free signal recovery is based on subspace detection (i.e., Grassmannian signaling), differential detection, and energy detection, respectively. For each approach, we review the theoretical foundation and design methodologies. We also provide comparative insights into their suitability across different channel models and system constraints, highlighting application scenarios where noncoherent methods offer performance and scalability advantages over traditional coherent communication. Furthermore, we discuss practical considerations of noncoherent communication, including compatibility with orthogonal frequency division multiplexing (OFDM), resilience to hardware impairments, and scalability with the number of users. Finally, we provide an outlook on future challenges and research directions in designin",
      "url": "https://openalex.org/W4416610237",
      "year": 2025,
      "venue": "arXiv (Cornell University)",
      "source": "openalex",
      "doi": "10.48550/arxiv.2505.23172",
      "pdf_url": "https://arxiv.org/pdf/2505.23172",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 301
    },
    {
      "title": "‘Integration-through-Law’: grand theory, revisionist history",
      "authors": [
        "Robert Schütze"
      ],
      "abstract": "Abstract How has the European Union been integrated in the past? Legal academics have traditionally pointed to the Court of Justice and to the broader idea of an ‘integration-through-law’. Through its supranational jurisprudence, the Court – not the EU legislature – was thus placed at the centre of the European integration project. The underlying reasons for this dominance of constitutional ‘law’ over legislative ‘politics’ have thereby been the subject of three famous explanations: the ‘equilibrium theory’ (Weiler), the ‘asymmetry theory’ (Scharpf) and the ‘over-constitutionalisation theory’ (Grimm). What are the merits of these grand theories of European integration when measured against the historical record? This article hopes to explore this question in the context of the internal market. Its historical revision begins with an analysis of the respective spheres of normative and decisional supranationalism during and after a foundational period (Sections 2 and 3). This is followed by an examination of the meaning and significance of the Cassis de Dijon judgment in the late 1970s. Through this revolutionary case, a dialectical relationship between the EU Court (‘law’) and the EU legislator (‘politics’) emerges (Section 4) that ultimately leads to the spectacular rise of EU legislation (Section 5) after the SEA. This transformational relationship will provide the critical lens for a historical revaluation of the three grand theories of legal integration (Section 6).",
      "url": "https://openalex.org/W4410790534",
      "year": 2025,
      "venue": "European Law Open",
      "source": "openalex",
      "doi": "10.1017/elo.2025.15",
      "pdf_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/F750BAFE5ECAC17C67D6667B85DCDFA1/S2752613525000153a.pdf/div-class-title-integration-through-law-grand-theory-revisionist-history-div.pdf",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 302
    },
    {
      "title": "Finding the right path: statistical mechanics of connected solutions in constraint satisfaction problems",
      "authors": [
        "Damien Barbier"
      ],
      "abstract": "We define and study a statistical mechanics ensemble that characterizes connected solutions in constraint satisfaction problems (CSPs). Built around a well-known local entropy bias, it allows us to better identify hardness transitions in problems where the energy landscape is dominated by isolated solutions. We apply this new device to the symmetric binary perceptron model (SBP), and study how its manifold of connected solutions behaves. We choose this particular problem because, while its typical solutions are isolated, it can be solved using local algorithms for a certain range of constraint density $α$ and threshold $κ$. With this new ensemble, we unveil the presence of a cluster composed of delocalized connected solutions. In particular, we demonstrate its stability until a critical threshold $κ^{\\rm no-mem}_{\\rm loc.\\, stab.}$ (dependent on $α$). This transition appears as paths of solutions shatter, a phenomenon that more conventional statistical mechanics approaches fail to grasp. Finally, we compared our predictions to simulations. For this, we used a modified Monte-Carlo algorithm, designed specifically to target these delocalized solutions. We obtained, as predicted, that the algorithm finds solutions until $κ\\approxκ^{\\rm no-mem}_{\\rm loc.\\, stab.}$.",
      "url": "https://openalex.org/W4415036511",
      "year": 2025,
      "venue": "arXiv (Cornell University)",
      "source": "openalex",
      "doi": "10.48550/arxiv.2505.20954",
      "pdf_url": "https://arxiv.org/pdf/2505.20954",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 303
    },
    {
      "title": "Learning Approaches to Dynamic Workflow Scheduling based on Genetic Programming and Deep Reinforcement Learning",
      "authors": [
        "Yifan Yang"
      ],
      "abstract": "&lt;p&gt;&lt;strong&gt;Dynamic workflow scheduling (DWS) in cloud computing is a critical yet challenging problem, involving assigning numerous workflow tasks to heterogeneous virtual machines under dynamic conditions to optimize cost or makespan. The complexity arises from unpredictable workflow arrivals and patterns, heterogeneous cloud resources, and rapidly evolving environments in workflow and resource status caused by real-time allocation. Among existing approaches, Priority Dispatching Rules (PDRs), are widely adopted for their intuitiveness, real-time efficiency, and ease of implementation. Manually designing effective PDRs is time-consuming, requires substantial domain expertise, and results in fixed structures that cannot adapt to various changes in dynamic environments. To address these limitations, this thesis develops advanced Genetic Programming Hyper-Heuristic (GPHH) and Deep Reinforcement Learning (DRL) approaches to automatically generate effective, generalizable, and adaptive PDRs for two practically important DWS problems, i.e., Deadline-Constrained Dynamic Workflow Scheduling in Cloud and Makespan-Aware Dynamic Workflow Scheduling in Cloud. Tree-based PDRs by GPHH have proven effective and interpretable in offline settings, avoiding the tediousness of manually designing rules. However, there is still room for improvement in terms of effectiveness and generalization. Neural network-based PDRs by DRL can address the adaptation challenge to rapidly changing e",
      "url": "https://openalex.org/W4410700106",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.26686/wgtn.29134007",
      "pdf_url": "https://openaccess.wgtn.ac.nz/articles/thesis/Learning_Approaches_to_Dynamic_Workflow_Scheduling_based_on_Genetic_Programming_and_Deep_Reinforcement_Learning/29134007/1/files/54766925.pdf",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 304
    },
    {
      "title": "Randomization Times under Quantum Chaotic Hamiltonian Evolution",
      "authors": [
        "Souradeep Ghosh",
        "Nicholas Hunter-Jones",
        "Joaquin F. Rodriguez-Nieva"
      ],
      "abstract": "Randomness generation through quantum-chaotic evolution underpins foundational questions in statistical mechanics and applications across quantum information science, including benchmarking, tomography, metrology, and demonstrations of quantum computational advantage. While statistical mechanics successfully captures the temporal averages of local observables, understanding randomness at the level of higher statistical moments remains a daunting challenge, with analytic progress largely confined to random quantum circuit models or fine-tuned systems exhibiting space-time duality. Here we study how much randomness can be dynamically generated by generic quantum-chaotic evolution under physical, non-random Hamiltonians. Combining theoretical insights with numerical simulations, we show that for broad classes of initially unentangled states, the dynamics become effectively Haar-random well before the system can ergodically explore the physically accessible Hilbert space. Both local and highly nonlocal observables, including entanglement measures, equilibrate to their Haar expectation values and fluctuations on polynomial timescales with remarkably high numerical precision, and with the fastest randomization occurring in regions of parameter space previously identified as maximally chaotic. Interestingly, this effective randomization can occur on timescales linear in system size, suggesting that the sub-ballistic growth of Renyi entropies typically observed in systems with conservation laws can be bypassed in non-random Hamiltonians with an appropriate choice of initial conditions.",
      "url": "http://arxiv.org/abs/2512.25074v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25074v1",
      "citations": null,
      "categories": [
        "cond-mat.stat-mech",
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 305
    },
    {
      "title": "GaMO: Geometry-aware Multi-view Diffusion Outpainting for Sparse-View 3D Reconstruction",
      "authors": [
        "Yi-Chuan Huang",
        "Hao-Jen Chien",
        "Chin-Yang Lin",
        "Ying-Huan Chen",
        "Yu-Lun Liu"
      ],
      "abstract": "Recent advances in 3D reconstruction have achieved remarkable progress in high-quality scene capture from dense multi-view imagery, yet struggle when input views are limited. Various approaches, including regularization techniques, semantic priors, and geometric constraints, have been implemented to address this challenge. Latest diffusion-based methods have demonstrated substantial improvements by generating novel views from new camera poses to augment training data, surpassing earlier regularization and prior-based techniques. Despite this progress, we identify three critical limitations in these state-of-the-art approaches: inadequate coverage beyond known view peripheries, geometric inconsistencies across generated views, and computationally expensive pipelines. We introduce GaMO (Geometry-aware Multi-view Outpainter), a framework that reformulates sparse-view reconstruction through multi-view outpainting. Instead of generating new viewpoints, GaMO expands the field of view from existing camera poses, which inherently preserves geometric consistency while providing broader scene coverage. Our approach employs multi-view conditioning and geometry-aware denoising strategies in a zero-shot manner without training. Extensive experiments on Replica and ScanNet++ demonstrate state-of-the-art reconstruction quality across 3, 6, and 9 input views, outperforming prior methods in PSNR and LPIPS, while achieving a $25\\times$ speedup over SOTA diffusion-based methods with processing time under 10 minutes. Project page: https://yichuanh.github.io/GaMO/",
      "url": "http://arxiv.org/abs/2512.25073v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25073v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 306
    },
    {
      "title": "Edit3r: Instant 3D Scene Editing from Sparse Unposed Images",
      "authors": [
        "Jiageng Liu",
        "Weijie Lyu",
        "Xueting Li",
        "Yejie Guo",
        "Ming-Hsuan Yang"
      ],
      "abstract": "We present Edit3r, a feed-forward framework that reconstructs and edits 3D scenes in a single pass from unposed, view-inconsistent, instruction-edited images. Unlike prior methods requiring per-scene optimization, Edit3r directly predicts instruction-aligned 3D edits, enabling fast and photorealistic rendering without optimization or pose estimation. A key challenge in training such a model lies in the absence of multi-view consistent edited images for supervision. We address this with (i) a SAM2-based recoloring strategy that generates reliable, cross-view-consistent supervision, and (ii) an asymmetric input strategy that pairs a recolored reference view with raw auxiliary views, encouraging the network to fuse and align disparate observations. At inference, our model effectively handles images edited by 2D methods such as InstructPix2Pix, despite not being exposed to such edits during training. For large-scale quantitative evaluation, we introduce DL3DV-Edit-Bench, a benchmark built on the DL3DV test split, featuring 20 diverse scenes, 4 edit types and 100 edits in total. Comprehensive quantitative and qualitative results show that Edit3r achieves superior semantic alignment and enhanced 3D consistency compared to recent baselines, while operating at significantly higher inference speed, making it promising for real-time 3D editing applications.",
      "url": "http://arxiv.org/abs/2512.25071v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25071v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 307
    },
    {
      "title": "Coordinated Humanoid Manipulation with Choice Policies",
      "authors": [
        "Haozhi Qi",
        "Yen-Jen Wang",
        "Toru Lin",
        "Brent Yi",
        "Yi Ma",
        "Koushil Sreenath",
        "Jitendra Malik"
      ],
      "abstract": "Humanoid robots hold great promise for operating in human-centric environments, yet achieving robust whole-body coordination across the head, hands, and legs remains a major challenge. We present a system that combines a modular teleoperation interface with a scalable learning framework to address this problem. Our teleoperation design decomposes humanoid control into intuitive submodules, which include hand-eye coordination, grasp primitives, arm end-effector tracking, and locomotion. This modularity allows us to collect high-quality demonstrations efficiently. Building on this, we introduce Choice Policy, an imitation learning approach that generates multiple candidate actions and learns to score them. This architecture enables both fast inference and effective modeling of multimodal behaviors. We validate our approach on two real-world tasks: dishwasher loading and whole-body loco-manipulation for whiteboard wiping. Experiments show that Choice Policy significantly outperforms diffusion policies and standard behavior cloning. Furthermore, our results indicate that hand-eye coordination is critical for success in long-horizon tasks. Our work demonstrates a practical path toward scalable data collection and learning for coordinated humanoid manipulation in unstructured environments.",
      "url": "http://arxiv.org/abs/2512.25072v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25072v1",
      "citations": null,
      "categories": [
        "cs.RO",
        "cs.AI",
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 308
    },
    {
      "title": "Classification of Interacting Topological Crystalline Superconductors in Three Dimensions and Beyond",
      "authors": [
        "Shang-Qiang Ning",
        "Xing-Yu Ren",
        "Qing-Rui Wang",
        "Yang Qi",
        "Zheng-Cheng Gu"
      ],
      "abstract": "Although classification for free-fermion topological superconductors (TSC) is established, systematically understanding the classification of 3D interacting TSCs remains difficult, especially those protected by crystalline symmetries like the 230 space groups. We build up a general framework for systematically classifying 3D interacting TSCs protected by crystalline symmetries together with discrete internal symmetries. We first establish a complete classification for fermionic symmetry protected topological phases (FSPT) with purely discrete internal symmetries, which determines the crystalline case via the crystalline equivalence principle. Using domain wall decoration, we obtain classification data and formulas for generic FSPTs, what are suitable for systematic computation. The four layers of decoration data $(n_1, n_2, n_3, ν_4)$ characterize a 3D FSPT with symmetry $G_b\\times_{ω_2}Z_2^f$, corresponding to $p+ip$, Kitaev chain, complex fermion, and bosonic SPT layers. Inspired by previous works, a crucial aspect is the $p+ip$ layer, where classification involves two possibilities: anti-unitary and infinite-order symmetries (e.g., translation). We show the former maps to some mirror FSPT classification with the mirror plane decorated by a $p+ip$ superconductor, while the latter is determined by the free part of $H^1(G_b, Z_T)$, corresponding to weak TSCs. Another key point is the Kitaev chain decoration for the anti-unitary symmetries, which differs essentially from unitary ones. We explicitly obtain formulas for all three layers of decoration $(n_2, n_3, ν_4)$, which are amenable to automatic computation. As an application, we classify the 230 space-group topological crystalline superconductors in interacting electronic systems.",
      "url": "http://arxiv.org/abs/2512.25069v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25069v1",
      "citations": null,
      "categories": [
        "cond-mat.str-el",
        "cond-mat.mes-hall",
        "cond-mat.supr-con"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 309
    },
    {
      "title": "No-cost Bell Nonlocality Certification from Quantum Tomography and Its Applications in Quantum Magic Witnessing",
      "authors": [
        "Pawel Cieslinski",
        "Lukas Knips",
        "Harald Weinfurter",
        "Wieslaw Laskowski"
      ],
      "abstract": "Tomographic measurements are the standard tool for characterizing quantum states, yet they are usually regarded only as means for state reconstruction or fidelity measurement. Here, we show that the same Pauli-basis measurements (X, Y, Z) can be directly employed for the certification of nonlocality at no additional experimental cost. Our framework allows any tomographic data - including archival datasets -- to be reinterpreted in terms of fundamental nonlocality tests. We introduce a generic, constructive method to generate tailored Bell inequalities and showcase their applicability to certify the non-locality of states in realistic experimental scenarios. Recognizing the stabilizer nature of the considered operators, we analyze our inequalities in the context of witnessing quantum magic - a crucial resource for quantum computing. Our approach requires Pauli measurements only and tests the quantum magic solely through the resources present in the state. Our results establish a universal standard that unifies state tomography with nonlocality certification and its application to quantum magic witnessing, thereby streamlining both fundamental studies and practical applications.",
      "url": "http://arxiv.org/abs/2512.25068v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25068v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 310
    },
    {
      "title": "FineTec: Fine-Grained Action Recognition Under Temporal Corruption via Skeleton Decomposition and Sequence Completion",
      "authors": [
        "Dian Shao",
        "Mingfei Shi",
        "Like Liu"
      ],
      "abstract": "Recognizing fine-grained actions from temporally corrupted skeleton sequences remains a significant challenge, particularly in real-world scenarios where online pose estimation often yields substantial missing data. Existing methods often struggle to accurately recover temporal dynamics and fine-grained spatial structures, resulting in the loss of subtle motion cues crucial for distinguishing similar actions. To address this, we propose FineTec, a unified framework for Fine-grained action recognition under Temporal Corruption. FineTec first restores a base skeleton sequence from corrupted input using context-aware completion with diverse temporal masking. Next, a skeleton-based spatial decomposition module partitions the skeleton into five semantic regions, further divides them into dynamic and static subgroups based on motion variance, and generates two augmented skeleton sequences via targeted perturbation. These, along with the base sequence, are then processed by a physics-driven estimation module, which utilizes Lagrangian dynamics to estimate joint accelerations. Finally, both the fused skeleton position sequence and the fused acceleration sequence are jointly fed into a GCN-based action recognition head. Extensive experiments on both coarse-grained (NTU-60, NTU-120) and fine-grained (Gym99, Gym288) benchmarks show that FineTec significantly outperforms previous methods under various levels of temporal corruption. Specifically, FineTec achieves top-1 accuracies of 89.1% and 78.1% on the challenging Gym99-severe and Gym288-severe settings, respectively, demonstrating its robustness and generalizability. Code and datasets could be found at https://smartdianlab.github.io/projects-FineTec/.",
      "url": "http://arxiv.org/abs/2512.25067v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25067v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 311
    },
    {
      "title": "From Inpainting to Editing: A Self-Bootstrapping Framework for Context-Rich Visual Dubbing",
      "authors": [
        "Xu He",
        "Haoxian Zhang",
        "Hejia Chen",
        "Changyuan Zheng",
        "Liyang Chen",
        "Songlin Tang",
        "Jiehui Huang",
        "Xiaoqiang Liu",
        "Pengfei Wan",
        "Zhiyong Wu"
      ],
      "abstract": "Audio-driven visual dubbing aims to synchronize a video's lip movements with new speech, but is fundamentally challenged by the lack of ideal training data: paired videos where only a subject's lip movements differ while all other visual conditions are identical. Existing methods circumvent this with a mask-based inpainting paradigm, where an incomplete visual conditioning forces models to simultaneously hallucinate missing content and sync lips, leading to visual artifacts, identity drift, and poor synchronization. In this work, we propose a novel self-bootstrapping framework that reframes visual dubbing from an ill-posed inpainting task into a well-conditioned video-to-video editing problem. Our approach employs a Diffusion Transformer, first as a data generator, to synthesize ideal training data: a lip-altered companion video for each real sample, forming visually aligned video pairs. A DiT-based audio-driven editor is then trained on these pairs end-to-end, leveraging the complete and aligned input video frames to focus solely on precise, audio-driven lip modifications. This complete, frame-aligned input conditioning forms a rich visual context for the editor, providing it with complete identity cues, scene interactions, and continuous spatiotemporal dynamics. Leveraging this rich context fundamentally enables our method to achieve highly accurate lip sync, faithful identity preservation, and exceptional robustness against challenging in-the-wild scenarios. We further introduce a timestep-adaptive multi-phase learning strategy as a necessary component to disentangle conflicting editing objectives across diffusion timesteps, thereby facilitating stable training and yielding enhanced lip synchronization and visual fidelity. Additionally, we propose ContextDubBench, a comprehensive benchmark dataset for robust evaluation in diverse and challenging practical application scenarios.",
      "url": "http://arxiv.org/abs/2512.25066v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25066v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 312
    },
    {
      "title": "The Logical Structure of Physical Laws: A Fixed Point Reconstruction",
      "authors": [
        "Eren Volkan Küçük"
      ],
      "abstract": "We formalise the self referential definition of physical laws using monotone operators on a lattice of theories, resolving the pathologies of naive set theoretic formulations. By invoking Tarski fixed point theorem, we identify physical theories as least fixed points of admissibility constraints derived from Galois connections. We demonstrate that QED and General Relativity can be represented in such a logical structure with respect to their symmetry and locality principles.",
      "url": "http://arxiv.org/abs/2512.25057v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25057v1",
      "citations": null,
      "categories": [
        "physics.hist-ph",
        "gr-qc",
        "math-ph",
        "math.LO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 313
    },
    {
      "title": "Sequential Bayesian parameter-state estimation in dynamical systems with noisy and incomplete observations via a variational framework",
      "authors": [
        "Liliang Wang",
        "Alex Gorodetsky"
      ],
      "abstract": "Online joint estimation of unknown parameters and states in a dynamical system with uncertainty quantification is crucial in many applications. For example, digital twins dynamically update their knowledge of model parameters and states to support prediction and decision-making. Reliability and computational speed are vital for DTs. Online parameter-state estimation ensures computational efficiency, while uncertainty quantification is essential for making reliable predictions and decisions. In parameter-state estimation, the joint distribution of the state and model parameters conditioned on the data, termed the joint posterior, provides accurate uncertainty quantification. Because the joint posterior is generally intractable to compute, this paper presents an online variational inference framework to compute its approximation at each time step. The approximation is factorized into a marginal distribution over the model parameters and a state distribution conditioned on the parameters. This factorization enables recursive updates through a two-stage procedure: first, the parameter posterior is approximated via variational inference; second, the state distribution conditioned on the parameters is computed using Gaussian filtering based on the estimated parameter posterior. The algorithmic design is supported by a theorem establishing upper bounds on the joint posterior approximation error. Numerical experiments demonstrate that the proposed method (i) matches the performance of the joint particle filter in low-dimensional problems, accurately inferring both unobserved states and unknown parameters of dynamical and observation models; (ii) remains robust under noisy, partial observations and model discrepancies in a chaotic Lorenz 96 system; and (iii) scales effectively to a high-dimensional convection-diffusion system, where it outperforms the joint ensemble Kalman filter.",
      "url": "http://arxiv.org/abs/2512.25056v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25056v1",
      "citations": null,
      "categories": [
        "stat.ME"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 314
    },
    {
      "title": "Context-aware LLM-based AI Agents for Human-centered Energy Management Systems in Smart Buildings",
      "authors": [
        "Tianzhi He",
        "Farrokh Jazizadeh"
      ],
      "abstract": "This study presents a conceptual framework and a prototype assessment for Large Language Model (LLM)-based Building Energy Management System (BEMS) AI agents to facilitate context-aware energy management in smart buildings through natural language interaction. The proposed framework comprises three modules: perception (sensing), central control (brain), and action (actuation and user interaction), forming a closed feedback loop that captures, analyzes, and interprets energy data to respond intelligently to user queries and manage connected appliances. By leveraging the autonomous data analytics capabilities of LLMs, the BEMS AI agent seeks to offer context-aware insights into energy consumption, cost prediction, and device scheduling, thereby addressing limitations in existing energy management systems. The prototype's performance was evaluated using 120 user queries across four distinct real-world residential energy datasets and different evaluation metrics, including latency, functionality, capability, accuracy, and cost-effectiveness. The generalizability of the framework was demonstrated using ANOVA tests. The results revealed promising performance, measured by response accuracy in device control (86%), memory-related tasks (97%), scheduling and automation (74%), and energy analysis (77%), while more complex cost estimation tasks highlighted areas for improvement with an accuracy of 49%. This benchmarking study moves toward formalizing the assessment of LLM-based BEMS AI agents and identifying future research directions, emphasizing the trade-off between response accuracy and computational efficiency.",
      "url": "http://arxiv.org/abs/2512.25055v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25055v1",
      "citations": null,
      "categories": [
        "cs.AI",
        "cs.HC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 315
    },
    {
      "title": "AdaGReS:Adaptive Greedy Context Selection via Redundancy-Aware Scoring for Token-Budgeted RAG",
      "authors": [
        "Chao Peng",
        "Bin Wang",
        "Zhilei Long",
        "Jinfang Sheng"
      ],
      "abstract": "Retrieval-augmented generation (RAG) is highly sensitive to the quality of selected context, yet standard top-k retrieval often returns redundant or near-duplicate chunks that waste token budget and degrade downstream generation. We present AdaGReS, a redundancy-aware context selection framework for token-budgeted RAG that optimizes a set-level objective combining query-chunk relevance and intra-set redundancy penalties. AdaGReS performs greedy selection under a token-budget constraint using marginal gains derived from the objective, and introduces a closed-form, instance-adaptive calibration of the relevance-redundancy trade-off parameter to eliminate manual tuning and adapt to candidate-pool statistics and budget limits. We further provide a theoretical analysis showing that the proposed objective exhibits epsilon-approximate submodularity under practical embedding similarity conditions, yielding near-optimality guarantees for greedy selection. Experiments on open-domain question answering (Natural Questions) and a high-redundancy biomedical (drug) corpus demonstrate consistent improvements in redundancy control and context quality, translating to better end-to-end answer quality and robustness across settings.",
      "url": "http://arxiv.org/abs/2512.25052v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25052v1",
      "citations": null,
      "categories": [
        "cs.CL",
        "cs.AI",
        "cs.IR"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 316
    },
    {
      "title": "The PDE-ODI principle and cylindrical mean curvature flows",
      "authors": [
        "Richard H. Bamler",
        "Yi Lai"
      ],
      "abstract": "We introduce a new approach for analyzing ancient solutions and singularities of mean curvature flow that are locally modeled on a cylinder. Its key ingredient is a general mechanism, called the \\emph{PDE--ODI principle}, which converts a broad class of parabolic differential equations into systems of ordinary differential inequalities. This principle bypasses many delicate analytic estimates used in previous work, and yields asymptotic expansions to arbitrarily high order.\n  As an application, we establish the uniqueness of the bowl soliton times a Euclidean factor among ancient, cylindrical flows with dominant linear mode. This extends previous results on this problem to the most general setting and is made possible by the stronger asymptotic control provided by our analysis. In the other case, when the quadratic mode dominates, we obtain a complete asymptotic expansion to arbitrary polynomial order, which will form the basis for a subsequent paper. Our framework also recovers and unifies several classical results. In particular, we give new proofs of the uniqueness of tangent flows (due to Colding-Minicozzi) and the rigidity of cylinders among shrinkers (due to Colding-Ilmanen-Minicozzi) by reducing both problems to a single ordinary differential inequality, without using the Łojasiewicz-Simon inequality.\n  Our approach is independent of prior work and the paper is largely self-contained.",
      "url": "http://arxiv.org/abs/2512.25050v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25050v1",
      "citations": null,
      "categories": [
        "math.DG",
        "math.AP"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 317
    },
    {
      "title": "Compound Estimation for Binomials",
      "authors": [
        "Yan Chen",
        "Lihua Lei"
      ],
      "abstract": "Many applications involve estimating the mean of multiple binomial outcomes as a common problem -- assessing intergenerational mobility of census tracts, estimating prevalence of infectious diseases across countries, and measuring click-through rates for different demographic groups. The most standard approach is to report the plain average of each outcome. Despite simplicity, the estimates are noisy when the sample sizes or mean parameters are small. In contrast, the Empirical Bayes (EB) methods are able to boost the average accuracy by borrowing information across tasks. Nevertheless, the EB methods require a Bayesian model where the parameters are sampled from a prior distribution which, unlike the commonly-studied Gaussian case, is unidentified due to discreteness of binomial measurements. Even if the prior distribution is known, the computation is difficult when the sample sizes are heterogeneous as there is no simple joint conjugate prior for the sample size and mean parameter.\n  In this paper, we consider the compound decision framework which treats the sample size and mean parameters as fixed quantities. We develop an approximate Stein's Unbiased Risk Estimator (SURE) for the average mean squared error given any class of estimators. For a class of machine learning-assisted linear shrinkage estimators, we establish asymptotic optimality, regret bounds, and valid inference. Unlike existing work, we work with the binomials directly without resorting to Gaussian approximations. This allows us to work with small sample sizes and/or mean parameters in both one-sample and two-sample settings. We demonstrate our approach using three datasets on firm discrimination, education outcomes, and innovation rates.",
      "url": "http://arxiv.org/abs/2512.25042v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25042v1",
      "citations": null,
      "categories": [
        "econ.EM",
        "math.ST",
        "stat.ME"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 318
    },
    {
      "title": "Perturbative Kondo destruction and global phase diagram of heavy fermion metals",
      "authors": [
        "Yiming Wang",
        "Shouvik Sur",
        "Chia-Chuan Liu",
        "Qimiao Si"
      ],
      "abstract": "Strange metals represent a foundational problem in quantum condensed matter physics, and heavy fermion systems provide a canonical setting to advance a general understanding. The concept of a Kondo destruction quantum critical point is widely invoked to describe the competition of the Kondo effect and the local-moment magnetism. Here, we develop a unified field-theoretic approach, analyzing this competition from a rare approach that is anchored by the magnetically ordered side. Our analysis reveals, for the first time within a renormalization group framework, a quantum critical point across which the Kondo effect goes from being destroyed to dominating. Our findings elucidate not only the Kondo destruction quantum criticality but also an accompanying global phase diagram of heavy fermion metals.",
      "url": "http://arxiv.org/abs/2512.25036v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25036v1",
      "citations": null,
      "categories": [
        "cond-mat.str-el"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 319
    },
    {
      "title": "Large Neutrino-Dark Matter Interactions: From Effective Field Theory to Ultraviolet Completions",
      "authors": [
        "K. S. Babu",
        "P. S. Bhupal Dev",
        "Anil Thapa"
      ],
      "abstract": "We develop a general effective field theory (EFT) framework for neutrino-dark matter (DM) interactions, and apply it to systematically find all possible gauge-invariant ultraviolet (UV) completions at a given EFT operator dimension. Our goal here is to find simple UV-complete models that can realize potentially large neutrino-DM interactions, while being consistent with all existing theoretical and experimental constraints. We first construct the leading non-derivative operator basis for neutrino-DM scattering in a low-energy effective theory with neutrinos and DM (DM-LEFT), together with its gauge-invariant embedding in the Standard Model EFT (DM-SMEFT). We then construct all renormalizable tree-level UV completions that generate the relevant DM-SMEFT operators up to dimension-8 using a topology-based classification. Using this framework, we present minimal UV-complete models for different DM types that can yield effective neutrino-DM couplings up to several orders of magnitude larger than the Fermi coupling, while satisfying all constraints, most notably from neutrino mass and from the charged-lepton sector. This includes a pseudo-Dirac fermion DM realization in the scotogenic neutrino mass model and models of Majorana DM inspired by type-II and inverse seesaw-based neutrino mass models. Phenomenological implications for DM thermal relic abundance and direct detection prospects, as well as various cosmological and laboratory constraints on the model parameter space, are also analyzed.",
      "url": "http://arxiv.org/abs/2512.25035v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25035v1",
      "citations": null,
      "categories": [
        "hep-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 320
    },
    {
      "title": "Universal Seesaw Pati-Salam Model with P for Strong CP",
      "authors": [
        "K. S. Babu",
        "Sumit Biswas"
      ],
      "abstract": "We develop a universal seesaw version of the Pati-Salam model wherein quarks and leptons of each family are unified into common multiplets transforming as $ψ_L(2,1,4))+ ψ_R((1,2,4)$ under the $SU(2)_L \\times SU(2)_R \\times SU(4)_c$ gauge symmetry. Parity symmetry is spontaneously broken in the model, which helps in solving the strong CP problem without the axion. The Higgs sector of the model is very simple, consisting of a single pair of $H_L(2,1,4)+ H_R(1,2,4)$ fields. Fermion masses arise through mixing of the chiral fermions with vector-like quarks and leptons contained in $(1,1,15)$ as well as $(1,1,10)_L+(1,1,10)_R$ multiplets via a universal seesaw mechanism. Consistency of such a spectrum with the observed quark and lepton masses is established. The parity solution to the strong CP problem is shown to be effective in this framework, although there are new loop contributions to $\\overlineθ$, compared to the analogous left-right symmetric model, arising from color sextet and octet fermions, as well as from diagrams mediated by leptoquark bosons. We also find that, in this setup, although lepton number is broken, neutrino masses remain zero at the tree-level. Small and finite Majorana neutrino masses are induced via one-loop diagrams, which we analyze and show to be compatible with oscillation experiments.",
      "url": "http://arxiv.org/abs/2512.25028v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25028v1",
      "citations": null,
      "categories": [
        "hep-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 321
    },
    {
      "title": "Computational Analysis of Disease Progression in Pediatric Pulmonary Arterial Hypertension",
      "authors": [
        "Omar Said",
        "Christopher Tossas-Betancourt",
        "Mary K. Olive",
        "Jimmy C. Lu",
        "Adam Dorfman",
        "C. Alberto Figueroa"
      ],
      "abstract": "Pulmonary arterial hypertension (PAH) is a progressive cardiopulmonary disease that leads to increased pulmonary pressures, vascular remodeling, and eventual right ventricular (RV) failure. Pediatric PAH remains understudied due to limited data and the lack of targeted diagnostic and therapeutic strategies. In this study, we developed and calibrated multi-scale, patient-specific cardiovascular models for four pediatric PAH patients using longitudinal MRI and catheterization data collected approximately two years apart. Using the CRIMSON simulation framework, we coupled three-dimensional fluid-structure interaction (FSI) models of the pulmonary arteries with zero-dimensional (0D) lumped-parameter heart and Windkessel models to simulate patient hemodynamics. An automated Python-based optimizer was developed to calibrate boundary conditions by minimizing discrepancies between simulated and clinical metrics, reducing calibration time from weeks to days. Model-derived metrics such as arterial stiffness, pulse wave velocity, resistance, and compliance were found to align with clinical indicators of disease severity and progression. Our findings demonstrate that computational modeling can non-invasively capture patient-specific hemodynamic adaptation over time, offering a promising tool for monitoring pediatric PAH and informing future treatment strategies.",
      "url": "http://arxiv.org/abs/2512.25027v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25027v1",
      "citations": null,
      "categories": [
        "physics.med-ph",
        "physics.comp-ph",
        "physics.flu-dyn"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 322
    },
    {
      "title": "Modewise Additive Factor Model for Matrix Time Series",
      "authors": [
        "Elynn Chen",
        "Yuefeng Han",
        "Jiayu Li",
        "Ke Xu"
      ],
      "abstract": "We introduce a Modewise Additive Factor Model (MAFM) for matrix-valued time series that captures row-specific and column-specific latent effects through an additive structure, offering greater flexibility than multiplicative frameworks such as Tucker and CP factor models. In MAFM, each observation decomposes into a row-factor component, a column-factor component, and noise, allowing distinct sources of variation along different modes to be modeled separately. We develop a computationally efficient two-stage estimation procedure: Modewise Inner-product Eigendecomposition (MINE) for initialization, followed by Complement-Projected Alternating Subspace Estimation (COMPAS) for iterative refinement. The key methodological innovation is that orthogonal complement projections completely eliminate cross-modal interference when estimating each loading space. We establish convergence rates for the estimated factor loading matrices under proper conditions. We further derive asymptotic distributions for the loading matrix estimators and develop consistent covariance estimators, yielding a data-driven inference framework that enables confidence interval construction and hypothesis testing. As a technical contribution of independent interest, we establish matrix Bernstein inequalities for quadratic forms of dependent matrix time series. Numerical experiments on synthetic and real data demonstrate the advantages of the proposed method over existing approaches.",
      "url": "http://arxiv.org/abs/2512.25025v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25025v1",
      "citations": null,
      "categories": [
        "stat.ME",
        "econ.EM",
        "math.ST"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 323
    },
    {
      "title": "MAMA-Memeia! Multi-Aspect Multi-Agent Collaboration for Depressive Symptoms Identification in Memes",
      "authors": [
        "Siddhant Agarwal",
        "Adya Dhuler",
        "Polly Ruhnke",
        "Melvin Speisman",
        "Md Shad Akhtar",
        "Shweta Yadav"
      ],
      "abstract": "Over the past years, memes have evolved from being exclusively a medium of humorous exchanges to one that allows users to express a range of emotions freely and easily. With the ever-growing utilization of memes in expressing depressive sentiments, we conduct a study on identifying depressive symptoms exhibited by memes shared by users of online social media platforms. We introduce RESTOREx as a vital resource for detecting depressive symptoms in memes on social media through the Large Language Model (LLM) generated and human-annotated explanations. We introduce MAMAMemeia, a collaborative multi-agent multi-aspect discussion framework grounded in the clinical psychology method of Cognitive Analytic Therapy (CAT) Competencies. MAMAMemeia improves upon the current state-of-the-art by 7.55% in macro-F1 and is established as the new benchmark compared to over 30 methods.",
      "url": "http://arxiv.org/abs/2512.25015v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25015v1",
      "citations": null,
      "categories": [
        "cs.CL"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 324
    },
    {
      "title": "Diffusion Language Models are Provably Optimal Parallel Samplers",
      "authors": [
        "Haozhe Jiang",
        "Nika Haghtalab",
        "Lijie Chen"
      ],
      "abstract": "Diffusion language models (DLMs) have emerged as a promising alternative to autoregressive models for faster inference via parallel token generation. We provide a rigorous foundation for this advantage by formalizing a model of parallel sampling and showing that DLMs augmented with polynomial-length chain-of-thought (CoT) can simulate any parallel sampling algorithm using an optimal number of sequential steps. Consequently, whenever a target distribution can be generated using a small number of sequential steps, a DLM can be used to generate the distribution using the same number of optimal sequential steps. However, without the ability to modify previously revealed tokens, DLMs with CoT can still incur large intermediate footprints. We prove that enabling remasking (converting unmasked tokens to masks) or revision (converting unmasked tokens to other unmasked tokens) together with CoT further allows DLMs to simulate any parallel sampling algorithm with optimal space complexity. We further justify the advantage of revision by establishing a strict expressivity gap: DLMs with revision or remasking are strictly more expressive than those without. Our results not only provide a theoretical justification for the promise of DLMs as the most efficient parallel sampler, but also advocate for enabling revision in DLMs.",
      "url": "http://arxiv.org/abs/2512.25014v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25014v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.CC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 325
    },
    {
      "title": "At the intersection of Numerical Analysis and Spectral Geometry",
      "authors": [
        "Nilima Nigam"
      ],
      "abstract": "How do the geometric properties of a domain impact the spectrum of an operator defined on it? How do we compute accurate and reliable approximations of these spectra? The former question is studied in spectral geometry, and the latter is a central concern in numerical analysis. In this short expository survey we revisit the process of eigenvalue approximation, from the perspective of computational spectral geometry. Over the years a multitude of methods -- for discretizing the operator and for the resultant discrete system -- have been developed and analyzed in the field of numerical analysis. High-accuracy and provably convergent discretization approaches can be used to examine the interplay between the spectrum of an operator and the geometric properties of the spatial domain or manifold it is defined on. While computations have been used to guide conjectures in spectral geometry, in recent years approximation-theoretic tools and validated computations are also being used as part of proof strategies in spectral geometry.\n  Given a particular spectral feature of interest, should we discretize the original problem, or seek a reformulation? Of the many possible approximation strategies, which should we choose? These choices are inextricably linked to the objective: on the one hand, rapid, specialized methods are often ideal for conjecture formulation (prioritizing efficiency and accuracy), whereas schemes with guaranteed, computable error bounds are needed when computation is incorporated into a proof strategy. We also review instances where the demanding requirements of spectral geometry -- the need for rigorous error control or the robust calculation of higher eigenvalues -- motivate new developments in numerical analysis.",
      "url": "http://arxiv.org/abs/2512.25012v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25012v1",
      "citations": null,
      "categories": [
        "math.NA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 326
    },
    {
      "title": "Uniqueness for stochastic differential equations in Hilbert spaces with irregular drift",
      "authors": [
        "Lukas Anzeletti",
        "Oleg Butkovsky",
        "Máté Gerencsér",
        "Alexander Shaposhnikov"
      ],
      "abstract": "We present a versatile framework to study strong existence and uniqueness for stochastic differential equations (SDEs) in Hilbert spaces with irregular drift. We consider an SDE in a separable Hilbert space $H$ \\begin{equation*} dX_t= (A X_t + b(X_t))dt +(-A)^{-γ/2}dW_t,\\quad X_0=x_0 \\in H, \\end{equation*} where $A$ is a self-adjoint negative definite operator with purely atomic spectrum, $W$ is a cylindrical Wiener process, $b$ is $α$-Hölder continuous function $H\\to H$, and a nonnegative parameter $γ$ such that the stochastic convolution takes values in $H$. We show that this equation has a unique strong solution provided that $α> 2γ/(1+γ)$. This substantially extends the seminal work of Da Prato and Flandoli (2010) as no structural assumption on $b$ is imposed. To obtain this result, we do not use infinite-dimensional Kolmogorov equations but instead develop a new technique combining Lê's theory of stochastic sewing in Hilbert spaces, Gaussian analysis, and a method of Lasry and Lions for approximation in Hilbert spaces.",
      "url": "http://arxiv.org/abs/2512.25003v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25003v1",
      "citations": null,
      "categories": [
        "math.PR",
        "math.AP"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 327
    },
    {
      "title": "The local limit of weighted spanning trees on balanced networks",
      "authors": [
        "Ágnes Kúsz"
      ],
      "abstract": "We prove that the local limit of the weighted spanning trees on any simple connected high degree almost regular sequence of electric networks is the Poisson(1) branching process conditioned to survive forever, by generalizing [NP22] and closing a gap in their proof. We also study the local statistics of the WST's on high degree almost balanced sequences, which is interesting even for the uniform spanning trees.\n  Our motivation comes from studying an interpolation $\\{\\mathsf{WST}^β(G)\\}_{β\\in [0, \\infty)}$ between UST(G) and MST(G) by WST's on a one-parameter family of random environments. This model has recently been introduced in [MSS24, Kús24], and the phases of several properties have been determined on the complete graphs.\n  We show a phase transition of $\\mathsf{WST}^{β_n}(G_n)$ regarding the local limit and expected edge overlaps for high degree almost balanced graph sequences $G_n$, without any structural assumptions on the graphs; while the expected total length is sensitive to the global structure of the graphs. Our general framework results in a better understanding even in the case of complete graphs, where it narrows the window of the phase transition of [Mak24].",
      "url": "http://arxiv.org/abs/2512.25001v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25001v1",
      "citations": null,
      "categories": [
        "math.PR",
        "math.CO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 328
    },
    {
      "title": "Bi-C2R: Bidirectional Continual Compatible Representation for Re-indexing Free Lifelong Person Re-identification",
      "authors": [
        "Zhenyu Cui",
        "Jiahuan Zhou",
        "Yuxin Peng"
      ],
      "abstract": "Lifelong person Re-IDentification (L-ReID) exploits sequentially collected data to continuously train and update a ReID model, focusing on the overall performance of all data. Its main challenge is to avoid the catastrophic forgetting problem of old knowledge while training on new data. Existing L-ReID methods typically re-extract new features for all historical gallery images for inference after each update, known as \"re-indexing\". However, historical gallery data typically suffers from direct saving due to the data privacy issue and the high re-indexing costs for large-scale gallery images. As a result, it inevitably leads to incompatible retrieval between query features extracted by the updated model and gallery features extracted by those before the update, greatly impairing the re-identification performance. To tackle the above issue, this paper focuses on a new task called Re-index Free Lifelong person Re-IDentification (RFL-ReID), which requires performing lifelong person re-identification without re-indexing historical gallery images. Therefore, RFL-ReID is more challenging than L-ReID, requiring continuous learning and balancing new and old knowledge in diverse streaming data, and making the features output by the new and old models compatible with each other. To this end, we propose a Bidirectional Continuous Compatible Representation (Bi-C2R) framework to continuously update the gallery features extracted by the old model to perform efficient L-ReID in a compatible manner. We verify our proposed Bi-C2R method through theoretical analysis and extensive experiments on multiple benchmarks, which demonstrate that the proposed method can achieve leading performance on both the introduced RFL-ReID task and the traditional L-ReID task.",
      "url": "http://arxiv.org/abs/2512.25000v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25000v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 329
    },
    {
      "title": "Basic Inequalities for First-Order Optimization with Applications to Statistical Risk Analysis",
      "authors": [
        "Seunghoon Paik",
        "Kangjie Zhou",
        "Matus Telgarsky",
        "Ryan J. Tibshirani"
      ],
      "abstract": "We introduce \\textit{basic inequalities} for first-order iterative optimization algorithms, forming a simple and versatile framework that connects implicit and explicit regularization. While related inequalities appear in the literature, we isolate and highlight a specific form and develop it as a well-rounded tool for statistical analysis. Let $f$ denote the objective function to be optimized. Given a first-order iterative algorithm initialized at $θ_0$ with current iterate $θ_T$, the basic inequality upper bounds $f(θ_T)-f(z)$ for any reference point $z$ in terms of the accumulated step sizes and the distances between $θ_0$, $θ_T$, and $z$. The bound translates the number of iterations into an effective regularization coefficient in the loss function. We demonstrate this framework through analyses of training dynamics and prediction risk bounds. In addition to revisiting and refining known results on gradient descent, we provide new results for mirror descent with Bregman divergence projection, for generalized linear models trained by gradient descent and exponentiated gradient descent, and for randomized predictors. We illustrate and supplement these theoretical findings with experiments on generalized linear models.",
      "url": "http://arxiv.org/abs/2512.24999v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24999v1",
      "citations": null,
      "categories": [
        "math.ST",
        "cs.LG",
        "math.NA",
        "math.OC",
        "stat.ML"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 330
    },
    {
      "title": "Numerical study of boson mixtures with multi-component continuous matrix product states",
      "authors": [
        "Wei Tang",
        "Benoît Tuybens",
        "Jutho Haegeman"
      ],
      "abstract": "The continuous matrix product state (cMPS) ansatz is a promising numerical tool for studying quantum many-body systems in continuous space. Although it provides a clean framework that allows one to directly simulate continuous systems, the optimization of cMPS is known to be a very challenging task, especially in the case of multi-component systems. In this work, we have developed an improved optimization scheme for multi-component cMPS that enables simulations of bosonic quantum mixtures with substantially larger bond dimensions than previous works. We benchmark our method on the two-component Lieb-Liniger model, obtaining numerical results that agree well with analytical predictions. Our work paves the way for further numerical studies of quantum mixture systems using the cMPS ansatz.",
      "url": "http://arxiv.org/abs/2512.24998v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24998v1",
      "citations": null,
      "categories": [
        "cond-mat.quant-gas",
        "cond-mat.str-el",
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 331
    },
    {
      "title": "Manifold classification from the descriptive viewpoint",
      "authors": [
        "Jeffrey Bergfalk",
        "Iian B. Smythe"
      ],
      "abstract": "We consider classification problems for manifolds and discrete subgroups of Lie groups from a descriptive set-theoretic point of view. This work is largely foundational in conception and character, recording both a framework for general study and Borel complexity computations for some of the most fundamental classes of manifolds. We show, for example, that for all $n\\geq 0$, the homeomorphism problem for compact topological $n$-manifolds is Borel equivalent to the relation $=_{\\mathbb{N}}$ of equality on the natural numbers, while the homeomorphism problem for noncompact topological $2$-manifolds is of maximal complexity among equivalence relations classifiable by countable structures. A nontrivial step in the latter consists of proving Borel measurable formulations of the Jordan--Schoenflies and surface triangulation theorems. Turning our attention to groups and geometric structures, we show, strengthening results of Stuck--Zimmer and Andretta--Camerlo--Hjorth, that the conjugacy relation on discrete subgroups of any noncompact semisimple Lie group is essentially countable universal. So too, as a corollary, is the isometry relation for complete hyperbolic $n$-manifolds for any $n\\geq 2$, generalizing a result of Hjorth--Kechris. We then show that the isometry relation for complete hyperbolic $n$-manifolds with finitely generated fundamental group is, in contrast, Borel equivalent to the equality relation $=_{\\mathbb{R}}$ on the real numbers when $n=2$, but that it is not concretely classifiable when $n=3$; thus there exists no Borel assignment of numerical complete invariants to finitely generated Kleinian groups up to conjugacy. We close with a survey of the most immediate open questions.",
      "url": "http://arxiv.org/abs/2512.24996v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24996v1",
      "citations": null,
      "categories": [
        "math.LO",
        "math.DG",
        "math.GT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 332
    },
    {
      "title": "Strong Gravitational Lensing by a Black Hole with a Global Monopole in Kalb-Ramond Bumblebee Gravity",
      "authors": [
        "Bijendra Kumar Vishvakarma",
        "Shubham Kala"
      ],
      "abstract": "We investigate the strong gravitational lensing and shadow properties of the black hole in the context of bumblebee gravity, characterized by a global monopole charge $κη^2$ and a Lorentz symmetry breaking parameter $γ$. We compute the deflection angles of light passing near the black hole in strong deflection limit, and estimate key lensing observables, including relativistic Einstein rings, absolute magnifications, image separations, and flux ratios, for astrophysical black holes. The black hole shadow is analyzed using the apparent angular size $θ_{\\rm Shadow} = 2\\,θ_{\\infty}$ in the limiting photon orbit. Furthermore, we study the modification of the shadow structure in the presence of a radially infalling, optically thin accretion flow within a generalized framework. Our results indicate that both the global monopole charge and Lorentz-violating parameters significantly influence the photon sphere, lensing observables, and shadow morphology, potentially providing observational signatures for testing bumblebee gravity in the strong-field regime.",
      "url": "http://arxiv.org/abs/2512.24995v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24995v1",
      "citations": null,
      "categories": [
        "gr-qc"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 333
    },
    {
      "title": "Mathieu Control of the Effective Coupling in Superconducting Qubits",
      "authors": [
        "Yi-Han Yu",
        "Xin-Yi Li",
        "Kai Xu",
        "Heng Fan"
      ],
      "abstract": "A common challenge in superconducting quantum circuits is the trade-off between strong coupling and computational subspace integrity. We present Mathieu control, which uses a non-resonant two-photon drive to create a selective nonlinear frequency shift. This shift modifies interactions while preserving qubit states, enabling continuous tuning of the ZZ coupling, including full suppression, and integrating single- and two-qubit gates with low leakage. For a qubit-coupler-qubit device, it allows independent ZZ control, facilitating a programmable Heisenberg (XXZ) Hamiltonian. Extended to a five-qubit chain, the system can be reconfigured to simulate dynamics of quantum magnetic phases. Mathieu control thus provides a framework for high-fidelity quantum logic and programmable simulation.",
      "url": "http://arxiv.org/abs/2512.24992v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24992v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 334
    },
    {
      "title": "Best Practices for Modelling Electrides",
      "authors": [
        "Lee A. Burton"
      ],
      "abstract": "Materials in which electrons occupy interstitial sites as anions are called electrides and exhibit unusual dimensionality-dependent electronic behavior. These properties make electrides attractive for catalysis, transparent conductors, and emergent quantum phenomena, yet their theoretical treatment remains challenging. In conventional materials, the ground-state atomic structure dictates the electronic configuration, whereas in electrides the electronic structure can instead govern the atomic arrangement. Here, the performance of commonly used exchange-correlation functionals is evaluated for representative one-, two-, and three-dimensional electrides. The results show that higher-cost approaches do not necessarily perform better across all cases, while standard methods capture the qualitative electride character and many key energetic and structural trends with surprising reliability. This behavior, likely arising from fortuitous error cancellation, supports the reliability of legacy studies in the field and the viability of efficient high-throughput exploration using low-cost methods. Overall, the findings support a tiered computational strategy for electride modelling, integrating system-specific heuristics with efficient first-principles screening. This approach balances computational feasibility with physical fidelity and underscores the continuing leadership of theory in the predictive discovery of electride materials across dimensionalities.",
      "url": "http://arxiv.org/abs/2512.24989v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24989v1",
      "citations": null,
      "categories": [
        "cond-mat.mtrl-sci"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 335
    },
    {
      "title": "PhysTalk: Language-driven Real-time Physics in 3D Gaussian Scenes",
      "authors": [
        "Luca Collorone",
        "Mert Kiray",
        "Indro Spinelli",
        "Fabio Galasso",
        "Benjamin Busam"
      ],
      "abstract": "Realistic visual simulations are omnipresent, yet their creation requires computing time, rendering, and expert animation knowledge. Open-vocabulary visual effects generation from text inputs emerges as a promising solution that can unlock immense creative potential. However, current pipelines lack both physical realism and effective language interfaces, requiring slow offline optimization. In contrast, PhysTalk takes a 3D Gaussian Splatting (3DGS) scene as input and translates arbitrary user prompts into real time, physics based, interactive 4D animations. A large language model (LLM) generates executable code that directly modifies 3DGS parameters through lightweight proxies and particle dynamics. Notably, PhysTalk is the first framework to couple 3DGS directly with a physics simulator without relying on time consuming mesh extraction. While remaining open vocabulary, this design enables interactive 3D Gaussian animation via collision aware, physics based manipulation of arbitrary, multi material objects. Finally, PhysTalk is train-free and computationally lightweight: this makes 4D animation broadly accessible and shifts these workflows from a \"render and wait\" paradigm toward an interactive dialogue with a modern, physics-informed pipeline.",
      "url": "http://arxiv.org/abs/2512.24986v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24986v1",
      "citations": null,
      "categories": [
        "cs.GR",
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 336
    },
    {
      "title": "Lindbladian PT phase transitions",
      "authors": [
        "Yuma Nakanishi",
        "Tomohiro Sasamoto"
      ],
      "abstract": "A parity-time (PT) transition is a spectral transition characteristic of non-Hermitian generators; it typically occurs at an exceptional point, where multiple eigenvectors coalesce. The concept of a PT transition has been extended to Markovian open quantum systems, which are described by the GKSL equation. Interestingly, the PT transition in many-body Markovian open quantum systems, the so-called \\textit{Lindbladian PT (L-PT) phase transition}, is closely related to two classes of exotic nonequilibrium many-body phenomena: \\textit{continuous-time crystals} and \\textit{non-reciprocal phase transitions}. In this review, we describe the recent advances in the study of L-PT phase transitions. First, we define PT symmetry in three distinct contexts: non-Hermitian systems, nonlinear dynamical systems, and Markovian open quantum systems, highlighting the interconnections between these frameworks. Second, we develop mean-field theories of L-PT phase transitions for collective-spin systems and for bipartite bosonic systems with particle-number conservation. Within these classes of models, we show that L-PT symmetry can induce a breaking of continuous time-translation symmetry down to a discrete one, leading to persistent periodic dynamics. We further demonstrate that the L-PT phase transition point is typically \\textit{a critical exceptional point}, where multiple collective excitation modes with zero excitation spectrum coalesce. These findings establish an explicit connection to continuous-time crystals and non-reciprocal phase transitions. Third, going beyond the mean-field theory, we analyze statistical and quantum properties, such as purity and quantum entanglement indicators of time-independent steady states for several specific models with the L-PT symmetry. Finally, we discuss future research directions for L-PT phase transitions.",
      "url": "http://arxiv.org/abs/2512.24981v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24981v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 337
    },
    {
      "title": "SymSeqBench: a unified framework for the generation and analysis of rule-based symbolic sequences and datasets",
      "authors": [
        "Barna Zajzon",
        "Younes Bouhadjar",
        "Maxime Fabre",
        "Felix Schmidt",
        "Noah Ostendorf",
        "Emre Neftci",
        "Abigail Morrison",
        "Renato Duarte"
      ],
      "abstract": "Sequential structure is a key feature of multiple domains of natural cognition and behavior, such as language, movement and decision-making. Likewise, it is also a central property of tasks to which we would like to apply artificial intelligence. It is therefore of great importance to develop frameworks that allow us to evaluate sequence learning and processing in a domain agnostic fashion, whilst simultaneously providing a link to formal theories of computation and computability. To address this need, we introduce two complementary software tools: SymSeq, designed to rigorously generate and analyze structured symbolic sequences, and SeqBench, a comprehensive benchmark suite of rule-based sequence processing tasks to evaluate the performance of artificial learning systems in cognitively relevant domains. In combination, SymSeqBench offers versatility in investigating sequential structure across diverse knowledge domains, including experimental psycholinguistics, cognitive psychology, behavioral analysis, neuromorphic computing and artificial intelligence. Due to its basis in Formal Language Theory (FLT), SymSeqBench provides researchers in multiple domains with a convenient and practical way to apply the concepts of FLT to conceptualize and standardize their experiments, thus advancing our understanding of cognition and behavior through shared computational frameworks and formalisms. The tool is modular, openly available and accessible to the research community.",
      "url": "http://arxiv.org/abs/2512.24977v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24977v1",
      "citations": null,
      "categories": [
        "q-bio.NC",
        "cs.AI",
        "cs.LG",
        "cs.NE"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 338
    },
    {
      "title": "Graphicality of power-law and double power-law degree sequences",
      "authors": [
        "Pietro Valigi",
        "M. Ángeles Serrano",
        "Claudio Castellano",
        "Lorenzo Cirigliano"
      ],
      "abstract": "The graphicality problem -- whether or not a sequence of integers can be used to create a simple graph -- is a key question in network theory and combinatorics, with many important practical applications. In this work, we study the graphicality of degree sequences distributed as a power-law with a size-dependent cutoff and as a double power-law with a size-dependent crossover. We combine the application of exact sufficient conditions for graphicality with heuristic conditions for nongraphicality which allow us to elucidate the physical reasons why some sequences are not graphical. For single power-laws we recover the known phase-diagram, we highlight the subtle interplay of distinct mechanisms violating graphicality and we explain why the infinite-size limit behavior is in some cases very far from being observed for finite sequences. For double power-laws we derive the graphicality of infinite sequences for all possible values of the degree exponents $γ_1$ and $γ_2$, uncovering a rich phase-diagram and pointing out the existence of five qualitatively distinct ways graphicality can be violated. The validity of theoretical arguments is supported by extensive numerical analysis.",
      "url": "http://arxiv.org/abs/2512.24976v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24976v1",
      "citations": null,
      "categories": [
        "cond-mat.dis-nn"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 339
    },
    {
      "title": "Hierarchical Deformation Planning and Neural Tracking for DLOs in Constrained Environments",
      "authors": [
        "Yunxi Tang",
        "Tianqi Yang",
        "Jing Huang",
        "Xiangyu Chu",
        "Kwok Wai Samuel Au"
      ],
      "abstract": "Deformable linear objects (DLOs) manipulation presents significant challenges due to DLOs' inherent high-dimensional state space and complex deformation dynamics. The wide-populated obstacles in realistic workspaces further complicate DLO manipulation, necessitating efficient deformation planning and robust deformation tracking. In this work, we propose a novel framework for DLO manipulation in constrained environments. This framework combines hierarchical deformation planning with neural tracking, ensuring reliable performance in both global deformation synthesis and local deformation tracking. Specifically, the deformation planner begins by generating a spatial path set that inherently satisfies the homotopic constraints associated with DLO keypoint paths. Next, a path-set-guided optimization method is applied to synthesize an optimal temporal deformation sequence for the DLO. In manipulation execution, a neural model predictive control approach, leveraging a data-driven deformation model, is designed to accurately track the planned DLO deformation sequence. The effectiveness of the proposed framework is validated in extensive constrained DLO manipulation tasks.",
      "url": "http://arxiv.org/abs/2512.24974v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24974v1",
      "citations": null,
      "categories": [
        "cs.RO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 340
    },
    {
      "title": "GEQIE Framework for Rapid Quantum Image Encoding",
      "authors": [
        "Rafał Potempa",
        "Michał Kordasz",
        "Józef P. Cyran",
        "Kamil Wereszczyński",
        "Krzysztof Simiński"
      ],
      "abstract": "This work presents a Python framework named after the General Equation of Quantum Image Encoding (GEQIE). The framework creates the image-encoding state using a unitary gate, which can later be transpiled to target quantum backends. The benchmarking results, simulated with different noise levels, demonstrate the correctness of the already implemented encoding methods and the usability of the framework for more sophisticated research tasks based on quantum image encodings. Additionally, we present a showcase example of Cosmic Web dark-matter density snapshot encoding and high-accuracy retrieval (PCC = 0.995) to demonstrate the extendability of the GEQIE framework to multidimensional data and its applicability to other fields of research.",
      "url": "http://arxiv.org/abs/2512.24973v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24973v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 341
    },
    {
      "title": "From Complex-Analytic Models to Sparse Domination: A Dyadic Approach of Hypersingular Operators via Bourgain's Interpolation Method",
      "authors": [
        "Bingyang Hu",
        "Xiaojing Zhou"
      ],
      "abstract": "Motivated by the work of Cheng--Fang--Wang--Yu on the hypersingular Bergman projection, we develop a real-variable and dyadic framework for hypersingular operators in regimes where strong-type estimates fail at the critical line. The main new input is a hypersingular sparse domination principle combined with Bourgain's interpolation method, which provides a flexible mechanism to establish critical-line (and endpoint) estimates.\n  In the unit disc setting with $1<t<3/2$, we obtain a full characterization of the $(p,q)$ mapping theory for the dyadic hypersingular maximal operator $\\mathcal M_t^{\\mathcal D}$, in particular including estimates on the critical line $1/q-1/p=2t-2$ and a weighted endpoint criterion in the radial setting. We also prove endpoint estimates for the hypersingular Bergman projection \\[ K_{2t}f(z)=\\int_{\\mathbb D}\\frac{f(w)}{(1-z\\overline w)^{2t}}\\,dA(w), \\] including a restricted weak-type bound at $(p,q)=\\bigl(\\tfrac{1}{3-2t},1\\bigr)$. Finally, we introduce a class of hypersingular cousin of sparse operators in $\\mathbb R^n$ associated with \\emph{graded} sparse families, quantified by the sparseness $η$ and a new structural parameter (the \\emph{degree}) $K_{\\mathcal S}$, and we characterize the corresponding strong/weak/restricted weak-type regimes in terms of $(n,t,η,K_{\\mathcal S})$.\n  Our real-variable perspective addresses to an inquiry raised by Cheng--Fang--Wang--Yu on developing effective real-analytic tools in the hypersingular regime for $K_{2t}$, and it also provides a new route toward the critical-line analysis of Forelli--Rudin type operators and related hypersingular operators in both real and complex settings.",
      "url": "http://arxiv.org/abs/2512.24972v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24972v1",
      "citations": null,
      "categories": [
        "math.CA",
        "math.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 342
    },
    {
      "title": "Random Batch Sum-of-Gaussians Method for Molecular Dynamics of Born-Mayer-Huggins Systems",
      "authors": [
        "Chen Chen",
        "Jiuyang Liang",
        "Zhenli Xu",
        "Qianru Zhang"
      ],
      "abstract": "The Born-Mayer-Huggins (BMH) potential, which combines Coulomb interactions with dispersion and short-range exponential repulsion, is widely used for ionic materials such as molten salts. However, large-scale molecular dynamics simulations of BMH systems are often limited by computation, communication, and memory costs. We recently proposed the random batch sum-of-Gaussians (RBSOG) method, which accelerates Coulomb calculations by using a sum-of-Gaussians (SOG) decomposition to split the potential into short- and long-range parts and by applying importance sampling in Fourier space for the long-range part. In this work, we extend the RBSOG to BMH systems and incorporate a random batch list (RBL) scheme to further accelerate the short-range part, yielding a unified framework for efficient simulations with the BMH potential. The combination of the SOG decomposition and the RBL enables an efficient and scalable treatment of both long- and short-range interactions in BMH system, particularly the RBL well handles the medium-range exponential repulsion and dispersion by the random batch neighbor list. Error estimate is provided to show the theoretical convergence of the RBL force. We evaluate the framework on molten NaCl and mixed alkali halide with up to $5\\times10^6$ atoms on $2048$ CPU cores. Compared to the Ewald-based particle-particle particle-mesh method and the RBSOG-only method, our method achieves approximately $4\\sim10\\times$ and $2\\times$ speedups while using $1000$ cores, respectively, under the same level of structural and thermodynamic accuracy and with a reduced memory usage. These results demonstrate the attractive performance of our method in accuracy and scalability for MD simulations with long-range interactions.",
      "url": "http://arxiv.org/abs/2512.24970v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24970v1",
      "citations": null,
      "categories": [
        "physics.comp-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 343
    },
    {
      "title": "Cosmic Himalayas in CROCODILE : Probing the Extreme Quasar Overdensities by Count-in-Cells analysis and Nearest Neighbor Distribution",
      "authors": [
        "Yuto Kuwayama",
        "Yongming Liang",
        "Kentaro Nagamine",
        "Yuri Oku",
        "Daisuke Nishihama",
        "Daisuke Toyouchi",
        "Keita Fukushima",
        "Hidenobu Yajima",
        "Hyunbae Park",
        "Masami Ouchi"
      ],
      "abstract": "The recently reported Cosmic Himalayas (CH) -- an extreme quasar overdensity at z~2 -- poses an apparent challenge to the Lambda CDM framework, with a reported significance of 16.9-sigma under Gaussian assumptions. Such an event appears improbably rare, with a formal probability of P ~ 10^-68. In this work, we investigate whether CH-like structures can naturally arise in cosmological hydrodynamic simulations. Using the CROCODILE simulation, which self-consistently models galaxy-black hole coevolution, we examine quasar clustering through two complementary approaches: the count-in-cells (CIC) statistic, which probes large-scale overdensities, and the nearest-neighbor distribution (NND), sensitive to small-scale environments. CIC analysis reveals that the underlying distribution is heavy-tailed and non-Gaussian, and that conventional Gaussian-based evaluation substantially overestimates the significance of extreme events. When modeled with an asymmetric generalized normal distribution (AGND), the inferred rarity of the CH is substantially reduced and reconciled with standard Lambda CDM; for instance, regions appearing as 12-sigma outliers under Gaussian assumptions (P ~ 10^-33) are found to occur in the AGND regime with a probability of P ~ 10^-4. NND analysis further demonstrates that extreme overdense regions within the simulation can naturally sustain two-point correlation function values similar to those observed in the CH (r0 ~ 30 Mpc/h), suggesting that the strong clustering stems from sample selection biases and local environmental variations. These two analyses conclusively highlight the importance of adopting non-Gaussian statistics when quantifying extreme overdensities of quasars and establish that the CH is not an anomaly, but a natural outcome of structure formation in the Lambda CDM universe.",
      "url": "http://arxiv.org/abs/2512.24966v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24966v1",
      "citations": null,
      "categories": [
        "astro-ph.CO",
        "astro-ph.GA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 344
    },
    {
      "title": "Approximating evolution operators of linear delay equations: a general framework for the convergence analysis",
      "authors": [
        "Alessia andò",
        "Giusy Bosco",
        "Dimitri Breda",
        "Davide Liessi"
      ],
      "abstract": "We consider the problem of discretizing evolution operators of linear delay equations with the aim of approximating their spectra, which is useful in investigating the stability properties of (nonlinear) equations via the principle of linearized stability. We develop a general convergence analysis based on a reformulation of the operators by means of a fixed-point equation, providing a list of hypotheses related to the regularization properties of the equation and the convergence of the chosen approximation techniques on suitable subspaces. This framework unifies the proofs for some methods based on pseudospectral discretization, which we present here in this new form. To exemplify the generality of the framework, we also apply it to a method of weighted residuals found in the literature, which was previously lacking a formal convergence analysis.",
      "url": "http://arxiv.org/abs/2512.24964v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24964v1",
      "citations": null,
      "categories": [
        "math.NA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 345
    },
    {
      "title": "From Principles to Effective Models: A Constructive Framework for Effective Covariant Actions with a Unique Vacuum Solution",
      "authors": [
        "Kristina Giesel",
        "Hongguang Liu"
      ],
      "abstract": "The absence of Birkhoff's theorem in effective quantum gravity models leads to a fundamental ambiguity in the vacuum sector, where a priori no unique vacuum solution exists. As a result, phenomenological investigations of the physical implications of these models have been made more difficult. We address this challenge by establishing a constructive framework which allows to formulate 4D covariant actions from the physical nature of the systems's degrees of freedom, which are dust and gravity, together with two guiding principles. We take advantage of the non-propagating nature of a relational dust clock and the suppression of gravitational waves in spherical symmetry. This structural ultralocality allows for a decomposition of the dynamics into independent LTB shells. We further impose spatial diffeomorphism invariance and a geometric guiding principle, where the latter ensures that a unique and static vacuum solution exists. These assumptions allow to strictly constrain the LTB shell Hamiltonian to a factorised form as well as the static vacuum metric function to a universal form. This constructive framework produces a fully 4D-covariant action that belongs to the class of generalised extended mimetic gravity models. This provides the necessary consistent basis for a perturbation theory in the context of quasi-normal modes or cosmological perturbations beyond the static sector in which quantum gravity effects are also included in linear and higher order perturbations. Furthermore, for this class of models our results resolve the long-standing `curvature polymerisation ambiguity' in loop quantum cosmology by unambiguously determining how flat space modifications are extended to non-flat geometries, thus unifying the description of black holes and cosmology in a single effective framework.",
      "url": "http://arxiv.org/abs/2512.24960v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24960v1",
      "citations": null,
      "categories": [
        "gr-qc",
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 346
    },
    {
      "title": "Semi-overlapping Multi-bandit Best Arm Identification for Sequential Support Network Learning",
      "authors": [
        "András Antos",
        "András Millinghoffer",
        "Péter Antal"
      ],
      "abstract": "Many modern AI and ML problems require evaluating partners' contributions through shared yet asymmetric, computationally intensive processes and the simultaneous selection of the most beneficial candidates. Sequential approaches to these problems can be unified under a new framework, Sequential Support Network Learning (SSNL), in which the goal is to select the most beneficial candidate set of partners for all participants using trials; that is, to learn a directed graph that represents the highest-performing contributions. We demonstrate that a new pure-exploration model, the semi-overlapping multi-(multi-armed) bandit (SOMMAB), in which a single evaluation provides distinct feedback to multiple bandits due to structural overlap among their arms, can be used to learn a support network from sparse candidate lists efficiently.\n  We develop a generalized GapE algorithm for SOMMABs and derive new exponential error bounds that improve the best known constant in the exponent for multi-bandit best-arm identification. The bounds scale linearly with the degree of overlap, revealing significant sample-complexity gains arising from shared evaluations.\n  From an application point of view, this work provides a theoretical foundation and improved performance guarantees for sequential learning tools for identifying support networks from sparse candidates in multiple learning problems, such as in multi-task learning (MTL), auxiliary task learning (ATL), federated learning (FL), and in multi-agent systems (MAS).",
      "url": "http://arxiv.org/abs/2512.24959v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24959v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 347
    },
    {
      "title": "AMAP Agentic Planning Technical Report",
      "authors": [
        "Yulan Hu",
        "Xiangwen Zhang",
        "Sheng Ouyang",
        "Hao Yi",
        "Lu Xu",
        "Qinglin Lang",
        "Lide Tan",
        "Xiang Cheng",
        "Tianchen Ye",
        "Zhicong Li"
      ],
      "abstract": "We present STAgent, an agentic large language model tailored for spatio-temporal understanding, designed to solve complex tasks such as constrained point-of-interest discovery and itinerary planning. STAgent is a specialized model capable of interacting with ten distinct tools within spatio-temporal scenarios, enabling it to explore, verify, and refine intermediate steps during complex reasoning. Notably, STAgent effectively preserves its general capabilities. We empower STAgent with these capabilities through three key contributions: (1) a stable tool environment that supports over ten domain-specific tools, enabling asynchronous rollout and training; (2) a hierarchical data curation framework that identifies high-quality data like a needle in a haystack, curating high-quality queries with a filter ratio of 1:10,000, emphasizing both diversity and difficulty; and (3) a cascaded training recipe that starts with a seed SFT stage acting as a guardian to measure query difficulty, followed by a second SFT stage fine-tuned on queries with high certainty, and an ultimate RL stage that leverages data of low certainty. Initialized with Qwen3-30B-A3B to establish a strong SFT foundation and leverage insights into sample difficulty, STAgent yields promising performance on TravelBench while maintaining its general capabilities across a wide range of general benchmarks, thereby demonstrating the effectiveness of our proposed agentic model.",
      "url": "http://arxiv.org/abs/2512.24957v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24957v1",
      "citations": null,
      "categories": [
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 348
    },
    {
      "title": "MSACL: Multi-Step Actor-Critic Learning with Lyapunov Certificates for Exponentially Stabilizing Control",
      "authors": [
        "Yongwei Zhang",
        "Yuanzhe Xing",
        "Quan Quan",
        "Zhikun She"
      ],
      "abstract": "Achieving provable stability in model-free reinforcement learning (RL) remains a challenge, particularly in balancing exploration with rigorous safety. This article introduces MSACL, a framework that integrates exponential stability theory with maximum entropy RL through multi-step Lyapunov certificate learning. Unlike methods relying on complex reward engineering, MSACL utilizes off-policy multi-step data to learn Lyapunov certificates satisfying theoretical stability conditions. By introducing Exponential Stability Labels (ESL) and a $λ$-weighted aggregation mechanism, the framework effectively balances the bias-variance trade-off in multi-step learning. Policy optimization is guided by a stability-aware advantage function, ensuring the learned policy promotes rapid Lyapunov descent. We evaluate MSACL across six benchmarks, including stabilization and nonlinear tracking tasks, demonstrating its superiority over state-of-the-art Lyapunov-based RL algorithms. MSACL achieves exponential stability and rapid convergence under simple rewards, while exhibiting significant robustness to uncertainties and generalization to unseen trajectories. Sensitivity analysis establishes the multi-step horizon $n=20$ as a robust default across diverse systems. By linking Lyapunov theory with off-policy actor-critic frameworks, MSACL provides a foundation for verifiably safe learning-based control. Source code and benchmark environments will be made publicly available.",
      "url": "http://arxiv.org/abs/2512.24955v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24955v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI",
        "cs.RO",
        "eess.SY"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 349
    },
    {
      "title": "VIPER: Process-aware Evaluation for Generative Video Reasoning",
      "authors": [
        "Yifan Li",
        "Yukai Gu",
        "Yingqian Min",
        "Zikang Liu",
        "Yifan Du",
        "Kun Zhou",
        "Min Yang",
        "Wayne Xin Zhao",
        "Minghui Qiu"
      ],
      "abstract": "Recent breakthroughs in video generation have demonstrated an emerging capability termed Chain-of-Frames (CoF) reasoning, where models resolve complex tasks through the generation of continuous frames. While these models show promise for Generative Video Reasoning (GVR), existing evaluation frameworks often rely on single-frame assessments, which can lead to outcome-hacking, where a model reaches a correct conclusion through an erroneous process. To address this, we propose a process-aware evaluation paradigm. We introduce VIPER, a comprehensive benchmark spanning 16 tasks across temporal, structural, symbolic, spatial, physics, and planning reasoning. Furthermore, we propose Process-outcome Consistency (POC@r), a new metric that utilizes VLM-as-Judge with a hierarchical rubric to evaluate both the validity of the intermediate steps and the final result. Our experiments reveal that state-of-the-art video models achieve only about 20% POC@1.0 and exhibit a significant outcome-hacking. We further explore the impact of test-time scaling and sampling robustness, highlighting a substantial gap between current video generation and true generalized visual reasoning. Our benchmark will be publicly released.",
      "url": "http://arxiv.org/abs/2512.24952v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24952v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 350
    },
    {
      "title": "The uncertainty constants: A unified framework of two, three and four observables",
      "authors": [
        "Minyi Huang"
      ],
      "abstract": "Uncertainty is a fundamental and important concept in quantum mechanics. Recent works have revealed both the product and sum forms of uncertainty constants for three observables. Such a result is intimately to the properties of Pauli operators. In this work, using the technique in matrix theory, we give an alternative proof for the case of three observables, and generalize the result to the case of four measurements. Comparing with the original proof, such a derivation is simplified. Moreover, the discussions can deal with the summation form of uncertainty relation for two, three and four observables in a unified way.",
      "url": "http://arxiv.org/abs/2512.24950v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24950v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 351
    },
    {
      "title": "ProDM: Synthetic Reality-driven Property-aware Progressive Diffusion Model for Coronary Calcium Motion Correction in Non-gated Chest CT",
      "authors": [
        "Xinran Gong",
        "Gorkem Durak",
        "Halil Ertugrul Aktas",
        "Vedat Cicek",
        "Jinkui Hao",
        "Ulas Bagci",
        "Nilay S. Shah",
        "Bo Zhou"
      ],
      "abstract": "Coronary artery calcium (CAC) scoring from chest CT is a well-established tool to stratify and refine clinical cardiovascular disease risk estimation. CAC quantification relies on the accurate delineation of calcified lesions, but is oftentimes affected by artifacts introduced by cardiac and respiratory motion. ECG-gated cardiac CTs substantially reduce motion artifacts, but their use in population screening and routine imaging remains limited due to gating requirements and lack of insurance coverage. Although identification of incidental CAC from non-gated chest CT is increasingly considered for it offers an accessible and widely available alternative, this modality is limited by more severe motion artifacts. We present ProDM (Property-aware Progressive Correction Diffusion Model), a generative diffusion framework that restores motion-free calcified lesions from non-gated CTs. ProDM introduces three key components: (1) a CAC motion simulation data engine that synthesizes realistic non-gated acquisitions with diverse motion trajectories directly from cardiac-gated CTs, enabling supervised training without paired data; (2) a property-aware learning strategy incorporating calcium-specific priors through a differentiable calcium consistency loss to preserve lesion integrity; and (3) a progressive correction scheme that reduces artifacts gradually across diffusion steps to enhance stability and calcium fidelity. Experiments on real patient datasets show that ProDM significantly improves CAC scoring accuracy, spatial lesion fidelity, and risk stratification performance compared with several baselines. A reader study on real non-gated scans further confirms that ProDM suppresses motion artifacts and improves clinical usability. These findings highlight the potential of progressive, property-aware frameworks for reliable CAC quantification from routine chest CT imaging.",
      "url": "http://arxiv.org/abs/2512.24948v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24948v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 352
    },
    {
      "title": "CPJ: Explainable Agricultural Pest Diagnosis via Caption-Prompt-Judge with LLM-Judged Refinement",
      "authors": [
        "Wentao Zhang",
        "Tao Fang",
        "Lina Lu",
        "Lifei Wang",
        "Weihe Zhong"
      ],
      "abstract": "Accurate and interpretable crop disease diagnosis is essential for agricultural decision-making, yet existing methods often rely on costly supervised fine-tuning and perform poorly under domain shifts. We propose Caption--Prompt--Judge (CPJ), a training-free few-shot framework that enhances Agri-Pest VQA through structured, interpretable image captions. CPJ employs large vision-language models to generate multi-angle captions, refined iteratively via an LLM-as-Judge module, which then inform a dual-answer VQA process for both recognition and management responses. Evaluated on CDDMBench, CPJ significantly improves performance: using GPT-5-mini captions, GPT-5-Nano achieves \\textbf{+22.7} pp in disease classification and \\textbf{+19.5} points in QA score over no-caption baselines. The framework provides transparent, evidence-based reasoning, advancing robust and explainable agricultural diagnosis without fine-tuning. Our code and data are publicly available at: https://github.com/CPJ-Agricultural/CPJ-Agricultural-Diagnosis.",
      "url": "http://arxiv.org/abs/2512.24947v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24947v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.CL"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 353
    },
    {
      "title": "HaineiFRDM: Explore Diffusion to Restore Defects in Fast-Movement Films",
      "authors": [
        "Rongji Xun",
        "Junjie Yuan",
        "Zhongjie Wang"
      ],
      "abstract": "Existing open-source film restoration methods show limited performance compared to commercial methods due to training with low-quality synthetic data and employing noisy optical flows. In addition, high-resolution films have not been explored by the open-source methods.We propose HaineiFRDM(Film Restoration Diffusion Model), a film restoration framework, to explore diffusion model's powerful content-understanding ability to help human expert better restore indistinguishable film defects.Specifically, we employ a patch-wise training and testing strategy to make restoring high-resolution films on one 24GB-VRAMR GPU possible and design a position-aware Global Prompt and Frame Fusion Modules.Also, we introduce a global-local frequency module to reconstruct consistent textures among different patches. Besides, we firstly restore a low-resolution result and use it as global residual to mitigate blocky artifacts caused by patching process.Furthermore, we construct a film restoration dataset that contains restored real-degraded films and realistic synthetic data.Comprehensive experimental results conclusively demonstrate the superiority of our model in defect restoration ability over existing open-source methods. Code and the dataset will be released.",
      "url": "http://arxiv.org/abs/2512.24946v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24946v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.AI",
        "cs.MM"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 354
    },
    {
      "title": "Inference for Delay Differential Equations Using Manifold-Constrained Gaussian Processes",
      "authors": [
        "Yuxuan Zhao",
        "S. W. Wong"
      ],
      "abstract": "Dynamic systems described by differential equations often involve feedback among system components. When there are time delays for components to sense and respond to feedback, delay differential equation (DDE) models are commonly used. This paper considers the problem of inferring unknown system parameters, including the time delays, from noisy and sparse experimental data observed from the system. We propose an extension of manifold-constrained Gaussian processes to conduct parameter inference for DDEs, whereas the time delay parameters have posed a challenge for existing methods that bypass numerical solvers. Our method uses a Bayesian framework to impose a Gaussian process model over the system trajectory, conditioned on the manifold constraint that satisfies the DDEs. For efficient computation, a linear interpolation scheme is developed to approximate the values of the time-delayed system outputs, along with corresponding theoretical error bounds on the approximated derivatives. Two simulation examples, based on Hutchinson's equation and the lac operon system, together with a real-world application using Ontario COVID-19 data, are used to illustrate the efficacy of our method.",
      "url": "https://www.semanticscholar.org/paper/59505531dc6e56a864c5a883b0860d09309620d3",
      "year": 2024,
      "venue": "Statistica sinica",
      "source": "semantic_scholar",
      "doi": "10.5705/ss.202024.0213",
      "pdf_url": "",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 355
    },
    {
      "title": "Absolute stability of program manifold of control systems with local connections",
      "authors": [
        "A. Tleulessova",
        "S. Zhumatov",
        "L. Zhapsarbayeva"
      ],
      "abstract": "The article presents the results carried out within the framework of the grant project of the Ministry of Science and Higher Education of the Republic of Kazakhstan AP19675193. In this paper we study the nonlinear problem of constructing material systems whose motions are described by ordinary differential equations, and the nonlinearity of the system satisfies the condition of a locally quadratic connection. We constructed the comparison systems to study stability in the neighborhood of program manifold. Moreover, we obtained the sufficient conditions for absolute stability with respect to a given function. Methods based on mappings of the state space of the original system into the state space of auxiliary comparison systems that preserve the dynamic properties under study are widely used in solving a manifold of theoretical and applied problems. In this work, matrix comparison equations are used to obtain sufficient stability conditions for the program manifold of non-autonomous control systems and systems with scalar control.",
      "url": "https://www.semanticscholar.org/paper/87ec5d9a59f19e26d8d9e2290682ffd040cc33bd",
      "year": 2024,
      "venue": "Bulletin of the National Engineering Academy of the Republic of Kazakhstan",
      "source": "semantic_scholar",
      "doi": "10.47533/2024.1606-146x.025",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 356
    },
    {
      "title": "Recurrent connections enable point attractor dynamics and dimensionality reduction in a connectome-constrained model of the insect learning center",
      "authors": [
        "Justin Joyce",
        "Raphael Norman-Tenazas",
        "Patricia K. Rivlin",
        "Grace M. Hwang",
        "Isaac Western",
        "Kechen Zhang",
        "William Gray-Roncal",
        "Brian Robinson"
      ],
      "abstract": "The learning center in the insect, the mushroom body (MB) with its predominant population of Kenyon Cells (KCs), is a widely studied model system to investigate neural processing principles, both experimentally and theoretically. While many computational models of the MB have been studied, the computational role of recurrent connectivity between KCs remains inadequately understood. Dynamical point attractors are a candidate theoretical framework where recurrent connections in a neural network can enable a discrete set of stable activation patterns. However, given that detailed, full recurrent connectivity patterns in biological neuron populations are mostly unknown, how theoretical models are substantiated by specific networks found in biology has not been clear. Leveraging the recent release of the full synapse-level connectivity of the MB in the fly, we performed a series of analyses and network model simulations to investigate the computational role of the recurrent KC connections, especially their significance in attractor dynamics. Structurally, the recurrent excitation (RE) connections are highly symmetric and balanced with feedforward input. In simulations, RE facilitates dimensionality reduction and allows a small set of self-sustaining point attractor states to emerge. To further quantify the possible range of network properties mediated by RE, we systematically explored the dynamical regimes enabled by changing recurrent connectivity strength. Finally, we establish connections between our findings and potential functional or behavioral implications. Overall, our work provides quantitative insights into the possible functional role of the recurrent excitatory connections in the MB by quantifying the point attractor network dynamics within a full synapse-level connectome-constrained highly recurrent network model. These findings advance our understanding of how biological neural networks may utilize point attractor dynamics. Author summary Point attractor neural networks are widely used theoretical models of associative memory, where recurrent connections between neurons enable a discrete set of stable activation patterns that can recover a full pattern based on partial cues. The detailed full recurrent connectivity patterns in biological neuron populations are largely unknown, however, raising questions about the precise correspondence between theoretical point attractor models and neural networks found in biology. Recent breakthroughs have unveiled the synapse-level connectivity of all neurons within the learning center of an insect, including recurrent connections between the primary neuron type—a crucial component with an elusive computational role. In this work, we perform analyses of these recurrent connectivity patterns and simulate neural network models that have these biologically constrained neural network patterns. We find that these recurrent connections are highly symmetric and balanced with input to the memory center. In simulations, we find that these recurrent connections perform dimensionality reduction and enable a small set of point attractor states. We additionally characterize how the strength of these recurrent connections affects network properties and downstream behavioral consequences. Overall, this work advances an understanding of the insect learning center as well as the relationship between theoretical and biological recurrent networks.",
      "url": "https://www.semanticscholar.org/paper/4877b7160bcde237a28d4448d0f7ad932ea919f0",
      "year": 2024,
      "venue": "bioRxiv",
      "source": "semantic_scholar",
      "doi": "10.1101/2024.01.10.574960",
      "pdf_url": "https://www.biorxiv.org/content/biorxiv/early/2024/01/11/2024.01.10.574960.full.pdf",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 357
    },
    {
      "title": "Gibbs Sampling from Human Feedback: A Provable KL- constrained Framework for RLHF",
      "authors": [
        "Wei Xiong",
        "Hanze Dong",
        "Chen Ye",
        "Han Zhong",
        "Nan Jiang",
        "Tong Zhang"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/e4435f282266da92d37066064c5239c6f96f0d64",
      "year": 2023,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2312.11456",
      "pdf_url": "",
      "citations": 65,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 358
    },
    {
      "title": "Hyper-Raman Spectra in Solution Based on the Reference Interaction Site Model Self-Consistent Field Method Coupled with the Constrained Spatial Electron Density Distribution and Vibrational Quasi-Degenerate Perturbation Theory.",
      "authors": [
        "K. Suda",
        "Kiyoshi Yagi",
        "Daisuke Yokogawa"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/2138811261773c1a6ce9c0c8f55d4178c18ed2ca",
      "year": 2025,
      "venue": "Journal of Physical Chemistry Letters",
      "source": "semantic_scholar",
      "doi": "10.1021/acs.jpclett.5c01961",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 359
    },
    {
      "title": "Quantum Manifold Optimization: A Design Framework for Future Communications Systems",
      "authors": [
        "Getuar Rexhepi",
        "Hyeon Seok Rou",
        "G. Abreu"
      ],
      "abstract": "Inspired by recent developments in various areas of science relevant to quantum computing, we introduce quantum manifold optimization (QMO) as a promising framework for solving constrained optimization problems in next-generation wireless communication systems. We begin by showing how classical wireless design problems - such as pilot design in cell-free (CF)-massive MIMO (mMIMO), beamformer optimization in gigantic multiple input multiple output (MIMO), and reconfigurable intelligent surface (RIS) phase tuning - naturally reside on structured manifolds like the Stiefel, Grassmannian, and oblique manifolds, with the latter novelly formulated in this work. Then, we demonstrate how these problems can be reformulated as trace-based quantum expectation values over variationally-encoded quantum states. While theoretical in scope, the work lays a foundation for a new class of quantum optimization algorithms with broad application to the design of future beyond-sixth-generation (B6G) systems.",
      "url": "https://www.semanticscholar.org/paper/aac0611bf6e1ae1009e228ec918070fc36afcec4",
      "year": 2025,
      "venue": "International Workshop on Signal Processing Advances in Wireless Communications",
      "source": "semantic_scholar",
      "doi": "10.1109/SPAWC66079.2025.11143361",
      "pdf_url": "",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 360
    },
    {
      "title": "Reconfiguring international manufacturing networks in times of uncertainty: towards a new theoretical framework",
      "authors": [
        "Eloi Letzelter",
        "Zheng Liu",
        "Yongjiang Shi",
        "Bo Yang"
      ],
      "abstract": "\n \n Research on multinational enterprises (MNEs) is now more relevant than ever, as the world economy experiences financial stress spilling over into the real sector from the credit-constrained private sector, large trade imbalances from low public saving and high volatility in energy prices. Concerns have emerged regarding supply chain resilience, prompting calls for MNEs to reconfigure their international manufacturing networks (IMNs) to respond to external disruptions. This article addresses two critical research questions: What is the impact of new external systems on IMNs? How can IMNs respond to the new external systems for dealing with the global economic crisis?\n \n \n \n The study adopts a qualitative case study approach by engaging with three MNEs based in Europe. Following a systematic thematic analysis process, we utilise a Context-Intervention-Mechanism-Outcome framework to develop our propositions.\n \n \n \n The article proposes a new theoretical framework to capture the interactions between IMNs and their external systems, namely, their international business environment. It provides six important IMN network reconfiguration mechanisms – E2E connection, vertical integration, localisation, diversification, simplification and streamlining, and building redundancy, alongside specific operating methods. The functions of the mechanisms result in enhanced IMN controllability, agility, resilience, and sustainability.\n \n \n \n In addition to providing important insights into the recent debate in the operations and supply chain management and international business literature. Our case-based approach offers the core guiding/controlling principles for IMNs to deal with uncertainty.\n",
      "url": "https://www.semanticscholar.org/paper/0e44eb72360f4ab0062c05ff5a93e01e3a7afb38",
      "year": 2025,
      "venue": "Business Process Management Journal",
      "source": "semantic_scholar",
      "doi": "10.1108/bpmj-07-2025-1100",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 361
    },
    {
      "title": "Affordances for throwing: An uncontrolled manifold analysis",
      "authors": [
        "Timothy Bennett",
        "Liam Thomas",
        "Andrew D Wilson"
      ],
      "abstract": "Movement systems are massively redundant, and there are always multiple movement solutions to any task demand; motor abundance. Movement consequently exhibits ‘repetition without repetition’, where movement outcomes are preserved but the kinematic details of the movement vary across repetitions. The uncontrolled manifold (UCM) concept is one of several methods that analyses movement variability with respect to task goals, to quantify repetition without repetition and test hypotheses about the control architecture producing a given abundant response to a task demand. However, like all these methods, UCM is under-constrained in how it decomposes a task and performance. In this paper, we propose and test a theoretical framework for constraining UCM analysis, specifically the perception of task-dynamical affordances. Participants threw tennis balls to hit a target set at 5m, 10m or 15m, and we performed UCM analysis on the shoulder-elbow-wrist joint angles with respect to variables derived from an affordance analysis of this task as well as more typical biomechanical variables. The affordance-based UCM analysis performed well, although data also showed thrower dynamics (effectivities) need to be accounted for as well. We discuss how the theoretical framework of affordances and affordance-based control can be connected to motor abundance methods in the future.",
      "url": "https://www.semanticscholar.org/paper/2b97cf191adab80d66e22bce114728b40e61a704",
      "year": 2024,
      "venue": "PLoS ONE",
      "source": "semantic_scholar",
      "doi": "10.1371/journal.pone.0301320",
      "pdf_url": "https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0301320&type=printable",
      "citations": 4,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 362
    },
    {
      "title": "Subspace-Constrained Quadratic Matrix Factorization: Algorithm and Applications",
      "authors": [
        "Zheng Zhai",
        "Xiaohui Li"
      ],
      "abstract": "Matrix Factorization has emerged as a widely adopted framework for modeling data exhibiting low-rank structures. To address challenges in manifold learning, this paper presents a subspace-constrained quadratic matrix factorization model. The model is designed to jointly learn key low-dimensional structures, including the tangent space, the normal subspace, and the quadratic form that links the tangent space to a low-dimensional representation. We solve the proposed factorization model using an alternating minimization method, involving an in-depth investigation of nonlinear regression and projection subproblems. Theoretical properties of the quadratic projection problem and convergence characteristics of the alternating strategy are also investigated. To validate our approach, we conduct numerical experiments on synthetic and real-world datasets. Results demonstrate that our model outperforms existing methods, highlighting its robustness and efficacy in capturing core low-dimensional structures.",
      "url": "https://www.semanticscholar.org/paper/6453ecb7dc383d3537438c939c70d95807e3cf63",
      "year": 2024,
      "venue": "Pattern Recognition",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2411.04717",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 363
    },
    {
      "title": "HySim-LLM: Embedding-Weighted Fine-Tuning Bounds and Manifold Denoising for Domain-Adapted LLMs",
      "authors": [
        "Majid Jaberi Douraki",
        "Hossein Sholehrasa",
        "Xuan Xu",
        "Remya Ampadi Ramachandran"
      ],
      "abstract": "The extraction and standardization of pharmacokinetic (PK) information from scientific literature remain significant challenges in computational pharmacology, which limits the reliability of data-driven models in drug development. Large language models (LLMs) have achieved remarkable progress in text understanding and reasoning, yet their adaptation to structured biomedical data, such as PK tables, remains constrained by heterogeneity, noise, and domain shift. To address these limitations, we propose HySim-LLM, a unified mathematical and computational framework that integrates embedding-weighted fine-tuning and manifold-aware denoising to enhance the robustness and interpretability of LLMs. We establish two theoretical results: (1) a similarity-weighted generalization bound that quantifies adaptation performance under embedding divergence, and (2) a manifold-based denoising guarantee that bounds loss contributions from noisy or off-manifold samples. These theorems provide a principled foundation for fine-tuning LLMs in structured biomedical settings. The framework offers a mathematically grounded pathway toward reliable and interpretable LLM adaptation for biomedical and data-intensive scientific domains.",
      "url": "https://www.semanticscholar.org/paper/3a54754f255ab932783e1f1b3508cd038c0cfa2d",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2510.07796",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 364
    },
    {
      "title": "GAMA: Geometry-Aware Manifold Alignment via Structured Adversarial Perturbations for Robust Domain Adaptation",
      "authors": [
        "Hana Satou",
        "F. Monkey"
      ],
      "abstract": "Domain adaptation remains a challenge when there is significant manifold discrepancy between source and target domains. Although recent methods leverage manifold-aware adversarial perturbations to perform data augmentation, they often neglect precise manifold alignment and systematic exploration of structured perturbations. To address this, we propose GAMA (Geometry-Aware Manifold Alignment), a structured framework that achieves explicit manifold alignment via adversarial perturbation guided by geometric information. GAMA systematically employs tangent space exploration and manifold-constrained adversarial optimization, simultaneously enhancing semantic consistency, robustness to off-manifold deviations, and cross-domain alignment. Theoretical analysis shows that GAMA tightens the generalization bound via structured regularization and explicit alignment. Empirical results on DomainNet, VisDA, and Office-Home demonstrate that GAMA consistently outperforms existing adversarial and adaptation methods in both unsupervised and few-shot settings, exhibiting superior robustness, generalization, and manifold alignment capability.",
      "url": "https://www.semanticscholar.org/paper/8f46ad9f995c5bec61085b5c36bf2d1b128be78d",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2505.15194",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 365
    },
    {
      "title": "Inequality Constraints on Statistical Submanifolds of Norden-Golden-like Statistical Manifold",
      "authors": [
        "Amit Kumar Rai",
        "Majid Ali Choudhary",
        "Mohammed Nisar",
        "F. Aloui"
      ],
      "abstract": "This paper explores novel inequalities for statistical submanifolds within the framework of the Norden golden-like statistical manifold. By leveraging the intrinsic properties of statistical manifolds and the structural richness of Norden golden geometry, we establish fundamental relationships between the intrinsic and extrinsic invariants of submanifolds. The methodology involves deriving generalized Chen-type and δ(2,2) curvature inequalities using curvature tensor analysis and dual affine connections. A concrete example is provided to verify the theoretical framework. The novelty of this work lies in extending classical curvature inequalities to a newly introduced statistical structure, thereby opening new perspectives in the study of geometric inequalities in information geometry and related mathematical physics contexts.",
      "url": "https://www.semanticscholar.org/paper/b3afb6b2a6315b04ffb2d764f653dbaaa327bb9a",
      "year": 2025,
      "venue": "Symmetry",
      "source": "semantic_scholar",
      "doi": "10.3390/sym17081206",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 366
    },
    {
      "title": "On the chemo-thermo-mechanics of constrained reactive mixtures of solids",
      "authors": [
        "Alberto Salvadori",
        "M. Serpelloni",
        "R. Mcmeeking"
      ],
      "abstract": "Building upon the classical chemo-mechanical theory of Larch{\\'e} and Cahn for equilibrium, numerous studies have investigated the transport of species in solids, with or without trapping phenomena. In most applications -- such as the swelling of hydrogels, hydrogen embrittlement in metals, and the transport of lithium or sodium in battery electrodes -- the formation of a new phase or compound can be directly associated with the concentration of the diffusing species. In the present work, we focus on the formation of solid mixtures made of multiple compounds, each characterized by its own volumetric expansion coefficient. Such a scenario arises, for instance, during the sodiation of tin anodes, among other systems. The classical chemo-mechanical framework is naturally recovered as a particular case of the proposed formulation. The theoretical framework developed herein elucidates and differentiates the concepts of phases and flowing species, while establishing rigorous connections between them. The present note is restricted to the general formulation of the governing equations, whereas application-specific developments will be addressed in forthcoming publications.",
      "url": "https://www.semanticscholar.org/paper/f719682c9b142da723f758856ffe0f47e9cb66db",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 367
    },
    {
      "title": "Information Geometry of the Retinal Representation Manifold",
      "authors": [
        "Xuehao Ding",
        "Dongsoo Lee",
        "Joshua B. Melander",
        "George Sivulka",
        "S. Ganguli",
        "S. Baccus"
      ],
      "abstract": "The ability for the brain to discriminate among visual stimuli is constrained by their retinal representations. Previous studies of visual discriminability have been limited to either low-dimensional artificial stimuli or pure theoretical considerations without a realistic encoding model. Here we propose a novel framework for understanding stimulus discriminability achieved by retinal representations of naturalistic stimuli with the method of information geometry. To model the joint probability distribution of neural responses conditioned on the stimulus, we created a stochastic encoding model of a population of salamander retinal ganglion cells based on a three-layer convolutional neural network model. This model not only accurately captured the mean response to natural scenes but also a variety of second-order statistics. With the model and the proposed theory, we computed the Fisher information metric over stimuli to study the most discriminable stimulus directions. We found that the most discriminable stimulus varied substantially across stimuli, allowing an examination of the relationship between the most discriminable stimulus and the current stimulus. By examining responses generated by the most discriminable stimuli we further found that the most discriminative response mode is often aligned with the most stochastic mode. This finding carries the important implication that under natural scenes, retinal noise correlations are information-limiting rather than increasing information transmission as has been previously speculated. We additionally observed that sensitivity saturates less in the population than for single cells and that as a function of firing rate, Fisher information varies less than sensitivity. We conclude that under natural scenes, population coding benefits from complementary coding and helps to equalize the information carried by different firing rates, which may facilitate decoding of the stimulus under principles of information maximization.",
      "url": "https://www.semanticscholar.org/paper/eeca6eb9453435f61e83744b1b135f2de9430ff7",
      "year": 2023,
      "venue": "bioRxiv",
      "source": "semantic_scholar",
      "doi": "10.1101/2023.05.17.541206",
      "pdf_url": "https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10245665",
      "citations": 13,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 368
    },
    {
      "title": "Isometric Manifold Learning Using Hierarchical Flow",
      "authors": [
        "Ziqi Pan",
        "Jianfu Zhang",
        "Li Niu",
        "Liqing Zhang"
      ],
      "abstract": "We propose the Hierarchical Flow (HF) model constrained by isometric regularizations for manifold learning that combines manifold learning goals such as dimensionality reduction, inference, sampling, projection and density estimation into one unified framework. Our proposed HF model is regularized to not only produce embeddings preserving the geometric structure of the manifold, but also project samples onto the manifold in a manner conforming to the rigorous definition of projection. Theoretical guarantees are provided for our HF model to satisfy the two desired properties. In order to detect the real dimensionality of the manifold, we also propose a two-stage dimensionality reduction algorithm, which is a time-efficient algorithm thanks to the hierarchical architecture design of our HF model. Experimental results justify our theoretical analysis, demonstrate the superiority of our dimensionality reduction algorithm in cost of training time, and verify the effect of the aforementioned properties in improving performances on downstream tasks such as anomaly detection.",
      "url": "https://www.semanticscholar.org/paper/558e966b86efc2ae667115c1f769bab5955c1a1a",
      "year": 2023,
      "venue": "AAAI Conference on Artificial Intelligence",
      "source": "semantic_scholar",
      "doi": "10.1609/aaai.v37i8.26124",
      "pdf_url": "https://ojs.aaai.org/index.php/AAAI/article/download/26124/25896",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 369
    },
    {
      "title": "Federated PCA on Grassmann Manifold for IoT Anomaly Detection",
      "authors": [
        "Tung-Anh Nguyen",
        "Long Tan Le",
        "Tuan Dung Nguyen",
        "Wei Bao",
        "Suranga Seneviratne",
        "Choong Seon Hong",
        "Nguyen H. Tran"
      ],
      "abstract": "With the proliferation of the Internet of Things (IoT) and the rising interconnectedness of devices, network security faces significant challenges, especially from anomalous activities. While traditional machine learning-based intrusion detection systems (ML-IDS) effectively employ supervised learning methods, they possess limitations such as the requirement for labeled data and challenges with high dimensionality. Recent unsupervised ML-IDS approaches such as AutoEncoders and Generative Adversarial Networks (GAN) offer alternative solutions but pose challenges in deployment onto resource-constrained IoT devices and in interpretability. To address these concerns, this paper proposes a novel federated unsupervised anomaly detection framework – FedPCA – that leverages Principal Component Analysis (PCA) and the Alternating Directions Method Multipliers (ADMM) to learn common representations of distributed non-i.i.d. datasets. Building on the FedPCA framework, we propose two algorithms, FedPE in Euclidean space and FedPG on Grassmann manifolds. Our approach enables real-time threat detection and mitigation at the device level, enhancing network resilience while ensuring privacy. Moreover, the proposed algorithms are accompanied by theoretical convergence rates even under a sub-sampling scheme, a novel result. Experimental results on the UNSW-NB15 and TON-IoT datasets show that our proposed methods offer performance in anomaly detection comparable to non-linear baselines, while providing significant improvements in communication and memory efficiency, underscoring their potential for securing IoT networks.",
      "url": "https://www.semanticscholar.org/paper/2008b835a5ad1beeb54e570306ce7946ff47f28d",
      "year": 2024,
      "venue": "IEEE/ACM Transactions on Networking",
      "source": "semantic_scholar",
      "doi": "10.1109/TNET.2024.3423780",
      "pdf_url": "https://arxiv.org/pdf/2407.07421",
      "citations": 16,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 370
    },
    {
      "title": "Complex Neutrosophic Soft Topology with Multivariate Analysis: A Unified Framework for Signal-Template Relationships",
      "authors": [
        "Maha Mohammed Saeed",
        "Raed Hatamleh",
        "Hamza Ali Abujabal",
        "Aqeedat Hussain",
        "Arif Mehmood Khattak",
        "M. Elashiry",
        "Abdelhalim Hasnaoui",
        "Alaa M. Abd El-latif"
      ],
      "abstract": "A novel theoretical method based on the topological formulation of complex neutrosophic soft sets (CNSS) is presented in this work. We go on to provide a thorough one-value complex neutrosophic soft topology, defining the most important topological properties such as interior, closure, and other related theoretical concerns, in order to give a strong mathematical foundation. An example is then provided to illustrate this frameworks analytical capabilities. To ensure a thorough empirical validation  of this paradigm, we employ a multifaceted approach that combines PCA, clustering, and manifold learning. Everyone agreed that a clear target hierarchy should be established, even though we employed a variety of analytical methodologies. In this hierarchy, T1 was consistently described as the most stable and representative target, T2 as a transitional outlier, and T3 as the most divergent target. In its most basic form, multivariate analysis and topological theory o er a versatile framework for clarifying the connections between templates and signals while avoiding the drawbacks of over-methodologization.",
      "url": "https://www.semanticscholar.org/paper/4e0f0c77e0055699928d5a30d4bf76fc53f639d1",
      "year": 2025,
      "venue": "European Journal of Pure and Applied Mathematics",
      "source": "semantic_scholar",
      "doi": "10.29020/nybg.ejpam.v18i4.7147",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 371
    },
    {
      "title": "Critical point search and linear response theory for computing electronic excitation energies of molecular systems. Part I: General framework, application to Hartree-Fock and DFT",
      "authors": [
        "L. Grazioli",
        "Yukuan Hu",
        "E. Cancès"
      ],
      "abstract": "Computing excited states of many-body quantum Hamiltonians is a fundamental challenge in computational physics and chemistry, with state-of-the-art methods broadly classified into variational (critical point search) and linear response approaches. The K\\\"ahler manifold formalism provides a uniform framework which naturally accommodates both strategies for a wide range of variational models, including Hartree-Fock, CASSCF, Full CI, and adiabatic TDDFT. In particular, this formalism leads to a systematic and straightforward way to obtain the final equations of linear response theory for nonlinear models, which provides, in the case of mean-field models (Hartree-Fock and DFT), a simple alternative to Casida's derivation. We detail the mathematical structure of Hamiltonian dynamics on K\\\"ahler manifolds, establish connections to standard quantum chemistry equations, and provide theoretical and numerical comparisons of excitation energy computation schemes at the Hartree-Fock level.",
      "url": "https://www.semanticscholar.org/paper/654be9b7f8e18c772e59ae429f472aa83660ef81",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 372
    },
    {
      "title": "RMVC: A Validated Algorithmic Framework for Decision-Making Under Uncertainty",
      "authors": [
        "Abdurrahman Dayıoğlu",
        "Fatma Ozen Erdogan",
        "B. Celik"
      ],
      "abstract": "The reliability of decision-making algorithms within soft set theory is fundamentally constrained by their underlying membership functions. Traditional binary approaches overlook the implicit connections between the attributes a candidate possesses and those it lacks—connections that can be inferred from the wider candidate pool. To address this core challenge, this paper puts forward the Relational Membership Value Calculation (RMVC), an algorithmic framework whose core is a fine-grained relational membership function. Our approach moves beyond binary logic to capture these nuanced interrelationships. We provide a rigorous theoretical analysis of the proposed algorithm, including its computational complexity and robustness, which is validated through a comprehensive sensitivity analysis. Crucially, a comparative analysis using the Gini Index quantitatively demonstrates that our method provides significantly higher granularity and discriminatory power on a representative case study. The RMVC is implemented as an open-source Python program, providing a foundational tool to enhance the reasoning capabilities of AI-driven decision support and expert systems.",
      "url": "https://www.semanticscholar.org/paper/577e6a637f1756e115bca38726d2f6476947e02f",
      "year": 2025,
      "venue": "Mathematics",
      "source": "semantic_scholar",
      "doi": "10.3390/math13162693",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 373
    },
    {
      "title": "A Riemannian gradient descent method for optimization on the indefinite Stiefel manifold",
      "authors": [
        "Dinh Van Tiep",
        "Nguyen Thanh Son"
      ],
      "abstract": "We consider the optimization problem with a generally quadratic matrix constraint of the form $X^TAX = J$, where $A$ is a given nonsingular, symmetric $n\\times n$ matrix and $J$ is a given $k\\times k$ symmetric matrix, with $k\\leq n$, satisfying $J^2 = I_k$. Since the feasible set constitutes a differentiable manifold, called the indefinite Stiefel manifold, we approach this problem within the framework of Riemannian optimization. Namely, we first equip the manifold with a Riemannian metric and construct the associated geometric structure, then propose a retraction based on the Cayley transform, and finally suggest a Riemannian gradient descent method using the attained materials, whose global convergence is guaranteed. Our results not only cover the known cases, the orthogonal and generalized Stiefel manifolds, but also provide a Riemannian optimization solution for other constrained problems which has not been investigated. As applications, we consider, via trace minimization, several eigenvalue problems of symmetric positive definite matrix pencils, including the linear response eigenvalue problem, and a matrix least square problem, a general framework for the Procrustes problem and constrained matrix equations. The presented numerical results justify the theoretical findings.",
      "url": "https://www.semanticscholar.org/paper/710dbd600ea6139bfc09db79096539f0ed2cebef",
      "year": 2024,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 374
    },
    {
      "title": "Towards a connection between the capacitated vehicle routing problem and the constrained centroid-based clustering",
      "authors": [
        "Abdelhakim Abdellaoui",
        "L. Benabbou",
        "Issmail Elhallaoui"
      ],
      "abstract": "Efficiently solving a vehicle routing problem (VRP) in a practical runtime is a critical challenge for delivery management companies. This paper explores both a theoretical and experimental connection between the Capacitated Vehicle Routing Problem (CVRP) and the Constrained Centroid-Based Clustering (CCBC). Reducing a CVRP to a CCBC is a synonym for a transition from an exponential to a polynomial complexity using commonly known algorithms for clustering, i.e K-means. At the beginning, we conduct an exploratory analysis to highlight the existence of such a relationship between the two problems through illustrative small-size examples and simultaneously deduce some mathematically-related formulations and properties. On a second level, the paper proposes a CCBC based approach endowed with some enhancements. The proposed framework consists of three stages. At the first step, a constrained centroid-based clustering algorithm generates feasible clusters of customers. This methodology incorporates three enhancement tools to achieve near-optimal clusters, namely: a multi-start procedure for initial centroids, a customer assignment metric, and a self-adjustment mechanism for choosing the number of clusters. At the second step, a traveling salesman problem (T SP) solver is used to optimize the order of customers within each cluster. Finally, we introduce a process relying on routes cutting and relinking procedure, which calls upon solving a linear and integer programming model to further improve the obtained routes. This step is inspired by the ruin&recreate algorithm. This approach is an extension of the classical cluster-first, route-second method and provides near-optimal solutions on well-known benchmark instances in terms of solution quality and computational runtime, offering a milestone in solving VRP.",
      "url": "https://www.semanticscholar.org/paper/4f97d67b87e9214723775a5fced8ee52fd95a07e",
      "year": 2024,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2403.14013",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 375
    },
    {
      "title": "Network Structures, Manifold Learning and Disease Management",
      "authors": [
        "J. V. Ramana raju",
        "R. Manjunatha"
      ],
      "abstract": "Infectious diseases continue to pose a significant global health challenge. Mathematical modeling has emerged as a critical tool for understanding disease dynamics and informing public health interventions. This paper provides a comprehensive overview of modelling infectious disease epidemics on networks, progressing from the simple Erdős-Rényi random graph to more complex structures. We delve into the intricate relationship between network topology and disease transmission in the setting of graph structures embedded on a manifold. A particular emphasis is placed on the connection between random graph theory, percolation theory, and dynamical systems, providing a robust theoretical framework for analyzing disease spread. Furthermore, the paper addresses the complexities introduced by networks with a high density of short closed loops, which can significantly impact disease dynamics. By examining these factors, we aim to contribute to the development of more accurate and effective models for predicting and controlling infectious disease outbreaks. We see a potential model to make informed decisions regarding targeted public health interventions, optimize resource allocation, and ultimately mitigate the impact of future epidemics.",
      "url": "https://www.semanticscholar.org/paper/3bec80b4f5e462cc19cff3bcb2fac545bc14f933",
      "year": 2025,
      "venue": "book.anvpublication",
      "source": "semantic_scholar",
      "doi": "10.52711/book.anv.icons-2024-006",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 376
    },
    {
      "title": "HardFlow: Hard-Constrained Sampling for Flow-Matching Models via Trajectory Optimization",
      "authors": [
        "Zeyang Li",
        "Kaveh Alim",
        "Navid Azizan"
      ],
      "abstract": "Diffusion and flow-matching have emerged as powerful methodologies for generative modeling, with remarkable success in capturing complex data distributions and enabling flexible guidance at inference time. Many downstream applications, however, demand enforcing hard constraints on generated samples (for example, robot trajectories must avoid obstacles), a requirement that goes beyond simple guidance. Prevailing projection-based approaches constrain the entire sampling path to the constraint manifold, which is overly restrictive and degrades sample quality. In this paper, we introduce a novel framework that reformulates hard-constrained sampling as a trajectory optimization problem. Our key insight is to leverage numerical optimal control to steer the sampling trajectory so that constraints are satisfied precisely at the terminal time. By exploiting the underlying structure of flow-matching models and adopting techniques from model predictive control, we transform this otherwise complex constrained optimization problem into a tractable surrogate that can be solved efficiently and effectively. Furthermore, this trajectory optimization perspective offers significant flexibility beyond mere constraint satisfaction, allowing for the inclusion of integral costs to minimize distribution shift and terminal objectives to further enhance sample quality, all within a unified framework. We provide a control-theoretic analysis of our method, establishing bounds on the approximation error between our tractable surrogate and the ideal formulation. Extensive experiments across diverse domains, including robotics (planning), partial differential equations (boundary control), and vision (text-guided image editing), demonstrate that our algorithm, which we name $\\textit{HardFlow}$, substantially outperforms existing methods in both constraint satisfaction and sample quality.",
      "url": "https://www.semanticscholar.org/paper/063901ad4d5725f17dff650c854497ec3517dd38",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 377
    },
    {
      "title": "A Fast Coordinate Descent Method for High-Dimensional Non-Negative Least Squares using a Unified Sparse Regression Framework",
      "authors": [
        "James Yang",
        "Trevor Hastie"
      ],
      "abstract": "We develop theoretical results that establish a connection across various regression methods such as the non-negative least squares, bounded variable least squares, simplex constrained least squares, and lasso. In particular, we show in general that a polyhedron constrained least squares problem admits a locally unique sparse solution in high dimensions. We demonstrate the power of our result by concretely quantifying the sparsity level for the aforementioned methods. Furthermore, we propose a novel coordinate descent based solver for NNLS in high dimensions using our theoretical result as motivation. We show through simulated data and a real data example that our solver achieves at least a 5x speed-up from the state-of-the-art solvers.",
      "url": "https://www.semanticscholar.org/paper/23ed07fa20720ba48e197c76f4a5502dc026e871",
      "year": 2024,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 378
    },
    {
      "title": "A Particle-Based Algorithm for Distributional Optimization on \\textit{Constrained Domains} via Variational Transport and Mirror Descent",
      "authors": [
        "Dai Hai Nguyen",
        "Tetsuya Sakurai"
      ],
      "abstract": "We consider the optimization problem of minimizing an objective functional, which admits a variational form and is defined over probability distributions on the constrained domain, which poses challenges to both theoretical analysis and algorithmic design. Inspired by the mirror descent algorithm for constrained optimization, we propose an iterative particle-based algorithm, named Mirrored Variational Transport (mirrorVT), extended from the Variational Transport framework [7] for dealing with the constrained domain. In particular, for each iteration, mirrorVT maps particles to an unconstrained dual domain induced by a mirror map and then approximately perform Wasserstein gradient descent on the manifold of distributions defined over the dual space by pushing particles. At the end of iteration, particles are mapped back to the original constrained domain. Through simulated experiments, we demonstrate the effectiveness of mirrorVT for minimizing the functionals over probability distributions on the simplex- and Euclidean ball-constrained domains. We also analyze its theoretical properties and characterize its convergence to the global minimum of the objective functional.",
      "url": "https://www.semanticscholar.org/paper/888d3af237f3bab73eb42ab66468c5284f00c450",
      "year": 2022,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.1007/s10994-023-06350-9",
      "pdf_url": "https://arxiv.org/pdf/2208.00587",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 379
    },
    {
      "title": "Stable Fine-Grained Image Synthesis with Gradient-Constrained Discriminators and Hierarchical Adaptive Sampling",
      "authors": [
        "Duo Wang",
        "Hengyi Wang"
      ],
      "abstract": "Recent advancements in conditional generative adversarial networks (GANs) have enabled fine-grained control over image attributes, yet critical challenges remain in long-tailed scenarios: unstable training dynamics from unbounded discriminator gradients and severe attribute entanglement caused by imbalanced distributions. A dual-mechanism framework integrating gradient-constrained discriminators and adaptive semantic balancing is proposed. First, a dynamic gradient penalty based on Lipschitz continuity constraints is developed to stabilize training by enforcing topological consistency between real and generated data manifolds. Second, a hierarchical conditional sampling strategy adaptively re-weights latent space distributions to mitigate local mode overfitting in imbalanced categories. Theoretical analysis reveals that unbounded discriminator gradients in traditional ACGANs lead to pattern degradation, while the proposed gradient constraints provably suppress this. Experiments on AnimeFace and CelebA-HQ demonstrate enhanced semantic controllability and visual consistency, particularly for rare attribute combinations. This work establishes a unified framework for stable fine-grained generation, advancing theoretical foundations in conditional GANs and practical applications in virtual content creation.",
      "url": "https://www.semanticscholar.org/paper/84dc3bf16c210ab3f1b60727b6b3e2b17e65faec",
      "year": 2025,
      "venue": "2025 6th International Conference on Computer Vision, Image and Deep Learning (CVIDL)",
      "source": "semantic_scholar",
      "doi": "10.1109/CVIDL65390.2025.11085797",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 380
    },
    {
      "title": "AdaProx: A Novel Method for Bilevel Optimization under Pessimistic Framework",
      "authors": [
        "Ziwei Guan",
        "Daouda Sow",
        "Sen-Fon Lin",
        "Yingbin Liang"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/ab514fc4be46ad64390072339be6c7b623603880",
      "year": 2025,
      "venue": "CPAL",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 381
    },
    {
      "title": "A Multi Affine Geometric Framework for Quantum Nonlocality. Unifying Berry Phases, Entanglement, and Coherence",
      "authors": [
        "Shoshauna Gauvin"
      ],
      "abstract": "We develop a multi affine geometric framework to unify classical and quantum mechanical laws through the lens of information geometry. By combining the principle of stationary action with maximum entropy production, we show that divergences in dual affine connections naturally give rise to quantum interference and nonlocal correlations. These ideas of maximum entropy and stationary align with the framework of entropic dynamics, wherein time evolution emerges from information-theoretic principles. A key insight comes from Berry phases. the mismatch of multiple affine connections leads to a nontrivial holonomy or\"area\", explaining how Bell type inequalities can be violated up to the Tsirelson bound. Additionally, we interpret sharply peaked distributions as\"pinning\"or\"sourcing\"curvature in the information manifold, drawing an analogy to a membrane under uniform stress. In this view, entanglement and coherence emerge as geometric features, phases and rotations within an overarching multi affine structure. We conclude that phenomena such as wavefunction collapse and quantum steering can be viewed as constraints on the manifold's\"laxity\"or curvature, with far-reaching implications for understanding quantum behavior in a geometric-information setting.",
      "url": "https://www.semanticscholar.org/paper/b1ed36496f10bfda1f0e9a9d65f6a23091f55523",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 382
    },
    {
      "title": "A note on the theoretical approach to Grassmannians and Pl\\\"ucker coordinates for additive skew-symmetric pairwise comparisons matrices",
      "authors": [
        "W. W. Koczkodaj",
        "Witold Pedrycz",
        "A. Pigazzini",
        "Laura P. Pigazzini",
        "R. Pinčák"
      ],
      "abstract": "Symmetry and antisymmetry are fundamental concepts in many strict sciences. Pairwise comparisons (PC) matrices are fundamental tools for representing pairwise relations in decision making. In this theoretical study, we present a novel framework that embeds additive skew-symmetric PC matrices into the Grassmannian manifold $G(2, n)$. This framework leverages Pl\\\"ucker coordinates to provide a rigorous geometric interpretation of their structure. Our key result demonstrates that the algebraic consistency condition $a_{ij} + a_{jk} - a_{ki} = 0$ is equivalent to the geometric consistency of $2$-planes in $G(2, n)$, satisfying the Pl\\\"ucker relations. This connection reveals that the algebraic properties of PC matrices can be naturally understood through their geometric representation. Additionally, we extend this framework by interpreting PC matrices as differential $2$-forms, providing a new perspective on their consistency as a closedness condition. Our framework of linear algebra, differential geometry, and algebraic geometry, placing PC matrices in a broader mathematical context. Rather than proposing a practical alternative to existing methods, our study aims to offer a theoretical foundation for future research by exploring new insights into higher-dimensional geometry and the geometric modeling of pairwise comparisons.",
      "url": "https://www.semanticscholar.org/paper/081a7e96b310870a7c268834ce4e4172e446b77e",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 383
    },
    {
      "title": "Dynamic consensus-building between neocortical areas via long-range connections",
      "authors": [
        "Mitra Javadzadeh",
        "Marine Schimel",
        "Sonja B. Hofer",
        "Yashar Ahmadian",
        "Guillaume Hennequin"
      ],
      "abstract": "The neocortex is organized into functionally specialized areas. While the functions and underlying neural circuitry of individual neocortical areas are well studied, it is unclear how these regions operate collectively to form percepts and implement cognitive processes. In particular, it remains unknown how distributed, potentially conflicting computations can be reconciled. Here we show that the reciprocal excitatory connections between cortical areas orchestrate neural dynamics to facilitate the gradual emergence of a ‘consensus’ across areas. We investigated the joint neural dynamics of primary (V1) and higher-order lateromedial (LM) visual areas in mice, using simultaneous multi-area electrophysiological recordings along with focal optogenetic perturbations to causally manipulate neural activity. We combined mechanistic circuit modeling with state-of-the-art data-driven nonlinear system identification, to construct biologically-constrained latent circuit models of the data that we could further interrogate. This approach revealed that long-range, reciprocal excitatory connections between V1 and LM implement an approximate line attractor in their joint dynamics, which promotes activity patterns encoding the presence of the stimulus consistently across the two areas. Further theoretical analyses revealed that the emergence of line attractor dynamics is a signature of a more general principle governing multi-area network dynamics: reciprocal inter-area excitatory connections reshape the dynamical landscape of the network, specifically slowing down the decay of activity patterns that encode stimulus features congruently across areas, while accelerating the decay of inconsistent patterns. This selective dynamic amplification leads to the emergence of multi-dimensional consensus between cortical areas about various stimulus features. Our analytical framework further predicted the timescales of specific activity patterns across areas, which we directly verified in our data. Therefore, by linking the anatomical organization of inter-area connections to the features they reconcile across areas, our work introduces a general theory of multi-area computation.",
      "url": "https://www.semanticscholar.org/paper/95de3fc684c5f98b85f85ee341d318e32c8d3d6d",
      "year": 2024,
      "venue": "bioRxiv",
      "source": "semantic_scholar",
      "doi": "10.1101/2024.11.27.625691",
      "pdf_url": "https://www.biorxiv.org/content/biorxiv/early/2024/11/27/2024.11.27.625691.full.pdf",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 384
    },
    {
      "title": "A convergence framework for energy minimisation of linear self-adjoint elliptic PDEs in nonlinear approximation spaces",
      "authors": [
        "Alexandre Magueresse",
        "Santiago Badia"
      ],
      "abstract": "Recent years have seen the emergence of nonlinear methods for solving partial differential equations (PDEs), such as physics-informed neural networks (PINNs). While these approaches often perform well in practice, their theoretical analysis remains limited, especially regarding convergence guarantees. This work develops a general optimisation framework for energy minimisation problems arising from linear self-adjoint elliptic PDEs, formulated over nonlinear but analytically tractable approximation spaces. The framework accommodates a natural split between linear and nonlinear parameters and supports hybrid optimisation strategies: linear variables are updated via linear solves or steepest descent, while nonlinear variables are handled using constrained projected descent. We establish both local and global convergence of the resulting algorithm under modular structural assumptions on the discrete energy functional, including differentiability, boundedness, regularity, and directional convexity. These assumptions are stated in an abstract form, allowing the framework to apply to a broad class of nonlinear approximation manifolds. In a companion paper [Magueresse, Badia (2025, arXiv:2508.17705)], we introduce a concrete instance of such a space based on overlapping free-knot tensor-product B-splines, which satisfies the required assumptions and enables geometrically adaptive solvers with rigorous convergence guarantees.",
      "url": "https://www.semanticscholar.org/paper/d9289d422ab129b2553f62761d60460784e9334e",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2508.17687",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 385
    },
    {
      "title": "Constrained Optimization with Compressed Gradients: A Dynamical Systems Perspective",
      "authors": [
        "Zhaoyue Xia",
        "Jun Du",
        "Chunxiao Jiang",
        "H. V. Poor",
        "Yong Ren"
      ],
      "abstract": "Gradient compression is of growing interests for solving constrained optimization problems including compressed sensing, noisy recovery and matrix completion under limited communication resources and storage costs. Convergence analysis of these methods from the dynamical systems viewpoint has attracted considerable attention because it provides a geometric demonstration towards the shadowing trajectory of a numerical scheme. In this work, we establish a tight connection between a continuous-time nonsmooth dynamical system called a perturbed sweeping process (PSP) and a projected scheme with compressed gradients. Theoretical results are obtained by analyzing the asymptotic pseudo trajectory of a PSP. We show that under mild assumptions a projected scheme converges to an internally chain transitive invariant set of the corresponding PSP. Furthermore, given the existence of a Lyapunov function $V$ with respect to a set $\\Lambda$, convergence to $\\Lambda$ can be established if $V(\\Lambda)$ has an empty interior. Based on these theoretical results, we are able to provide a useful framework for convergence analysis of projected methods with compressed gradients. Moreover, we propose a provably convergent distributed compressed gradient descent algorithm for distributed nonconvex optimization. Finally, numerical simulations are conducted to confirm the validity of theoretical analysis and the effectiveness of the proposed algorithm.",
      "url": "https://www.semanticscholar.org/paper/6e94041cc04a48e3b29161476e78c9641a7cd0b5",
      "year": 2024,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 386
    },
    {
      "title": "Constrained Nonlinear Output Regulation Using Model Predictive Control",
      "authors": [
        "Johannes Köhler",
        "M. Müller",
        "F. Allgöwer"
      ],
      "abstract": "We present a model predictive control (MPC) framework to solve the constrained nonlinear output regulation problem. The main feature of the proposed framework is that the application does not require the solution to classical regulator (Francis–Byrnes–Isidori) equations or any other offline design procedure. In particular, the proposed formulation simply minimizes the predicted output error, possibly with some input regularization. Instead of using terminal cost/sets or a positive-definite stage cost as is standard in MPC theory, we build on the theoretical results by Grimm et al. using a detectability notion. The proposed formulation is applicable if the constrained nonlinear regulation problem is (strictly) feasible; the plant is incrementally stabilizable and incrementally input–output to state stable (i-IOSS, detectable). We show that for minimum phase systems, such a design ensures exponential stability of the regulator manifold. We also provide a design procedure in case of unstable zero dynamics using an incremental input regularization and a nonresonance condition. The theoretical results are illustrated with an example involving offset-free tracking.",
      "url": "https://www.semanticscholar.org/paper/38c1261f329dda0c799e45f450d40d8b44347f9e",
      "year": 2020,
      "venue": "IEEE Transactions on Automatic Control",
      "source": "semantic_scholar",
      "doi": "10.1109/TAC.2021.3081080",
      "pdf_url": "https://arxiv.org/pdf/2005.12413",
      "citations": 28,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 387
    },
    {
      "title": "Dynamic Brain Network Modeling Based on Nonlinear Low-Rank Manifold Regularization",
      "authors": [
        "San-Wang Wang",
        "Zhigang Li",
        "Kai Yuan",
        "Shan-Shan Qu",
        "Xin Wen",
        "Wei-Feng Mi",
        "Qunxi Dong",
        "Lin Lu",
        "Gaohua Wang"
      ],
      "abstract": "Functional brain network modeling plays a crucial role in uncovering cognitive mechanisms and identifying abnormalities associated with brain disorders. However, traditional approaches—such as Pearson correlation and mutual information—typically assume that interregional relationships are static and linear, limiting their ability to capture dynamic interactions and nonlinear features. To address this limitation, we propose a nonlinear dynamic low-rank representation method. This approach constructs interregional similarity weights using physiological state matrices and Frobenius distance, while incorporating neighborhood information to generate a row-stochastic transition matrix, thereby enhancing the robustness of local connections. Additionally, a dynamic interaction effect matrix is constructed based on the stationary distribution eigenvectors, enabling the identification of both direct and indirect information transmission processes between brain regions. The method preserves the brain’s modular structure through low-rank representation and manifold regularization. Experimental results demonstrate that the proposed method not only effectively reveals dynamic information transmission pathways, cross-modular cooperative effects, and task-dependent hub reorganization patterns, but also significantly outperforms traditional static connectivity approaches. This study offers a mathematically rigorous and physiologically interpretable framework for dynamic brain network modeling and lays a theoretical foundation for the detection of dynamic abnormalities in mental disorders.",
      "url": "https://www.semanticscholar.org/paper/21d8ac40aa38c5f1128bfc515d5181e82faa7604",
      "year": 2025,
      "venue": "IEEE Transactions on Computational Social Systems",
      "source": "semantic_scholar",
      "doi": "10.1109/TCSS.2025.3573770",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 388
    },
    {
      "title": "Complex harmonics reveal low-dimensional manifolds of critical brain dynamics",
      "authors": [
        "G. Deco",
        "Y. Perl",
        "M. Kringelbach"
      ],
      "abstract": "The brain needs to perform time-critical computations to ensure survival. A potential solution lies in the non-local, distributed computation at the whole-brain level made possible by criticality and amplified by the rare long-range connections found in the brain’s unique anatomical structure. This non-locality can be captured by the mathematical structure of Schrödinger’s wave equation, which is at the heart of the novel CHARM (Complex Harmonics decomposition) framework that performs the necessary dimensional manifold reduction able to extract non-locality in critical spacetime brain dynamics. Using a large neuroimaging dataset of over 1,000 people, CHARM captured the critical, non-local/long-range nature of brain dynamics and the underlying mechanisms were established using a precise whole-brain model. Equally, CHARM revealed the significantly different critical dynamics of wakefulness and sleep. Overall, CHARM is a promising theoretical framework for capturing the low-dimensionality of the complex network dynamics observed in neuroscience and provides evidence that networks of brain regions rather than individual brain regions are the key computational engines of critical brain dynamics.",
      "url": "https://www.semanticscholar.org/paper/79c92bde562ae0356acda295ead7128e8c11928b",
      "year": 2024,
      "venue": "bioRxiv",
      "source": "semantic_scholar",
      "doi": "10.1101/2024.06.15.599165",
      "pdf_url": "https://www.biorxiv.org/content/biorxiv/early/2024/06/16/2024.06.15.599165.full.pdf",
      "citations": 10,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 389
    },
    {
      "title": "Unfolding the Manifold Flavours of Causality",
      "authors": [
        "Rui A. P. Perdigão"
      ],
      "abstract": "The present work provides a contribution to an overarching cross-methodological causality investigation, encompassing a methodological synergy among physical, analytical, information-theoretic and systems intelligence approaches to causal discovery and quantification in complex system dynamics. These efforts methodologically lead to the emergence of a broader causal framework, valid not only in classical recurrence-based dynamical systems, but also on the generalized information physics of non-ergodic coevolutionary spatiotemporal complexity. This study begins with a comprehensive cross-examination of causality metrics derived from these diverse domains, by synthesizing causality insights from information theory, which enables the quantification of information flow among variables; differential geometry, which captures the curvature and structure of causal relationships; dynamical systems, which analyze the temporal evolution of systems and associated kinematic geometric properties; and fundamental physical metrics, which elucidate causal connections in the physical world from fundamental thermodynamic principles. Through this analysis, we aim to deepen our understanding of causality in complex systems, with physical process understanding and geophysical applications in mind. Having laid out some of the key methodological flavours of causality, the present communication introduces new metrics further contributing to a broader non-Shannonian information theoretic causality pool of methods, along with additional advances on quantum thermodymamical, nonlinear statistical mechanical, differential geometric and topologic approaches on causality. Positioning ourselves in a broader nonlinear non-Gaussian non-ergodic setting to tackle far-from-equilibrium structural-functional coevolution and synergistic emergence in complex system dynamics, our derivations further contribute to a new generation of information theoretic, dynamical systems and non-equilibrium thermodynamic causality approaches, along with their synergistic articulation towards a unified framework. This brings out further cross-methodological comparability, portability and complementary insights on dealing with the intricate causality of complex multiscale far-from-equilibrium Earth system dynamic phenomena. By unveiling manifold flavours of causality and their interconnections, this study brings out their commonalities, synergies, and further potential synergistic applications across disciplines. This interdisciplinary approach not only enhances our theoretical understanding of causality but also provides practical implications for applications in fields such as data science, network theory, and complex systems analysis, with direct relevance across a plethora of scientific, technical and operational fields.",
      "url": "https://www.semanticscholar.org/paper/ddb8ffb2f5fd9c939b35034b7206fae818eae999",
      "year": 2024,
      "venue": "",
      "source": "semantic_scholar",
      "doi": "10.46337/mdsc.1804",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 390
    },
    {
      "title": "Generalized Dimension Reduction Using Semi-Relaxed Gromov-Wasserstein Distance",
      "authors": [
        "Ranthony A. Clark",
        "Tom Needham",
        "Thomas Weighill"
      ],
      "abstract": "Dimension reduction techniques typically seek an embedding of a high-dimensional point cloud into a low-dimensional Euclidean space which optimally preserves the geometry of the input data. Based on expert knowledge, one may instead wish to embed the data into some other manifold or metric space in order to better reflect the geometry or topology of the point cloud. We propose a general method for manifold-valued multidimensional scaling based on concepts from optimal transport. In particular, we establish theoretical connections between the recently introduced semi-relaxed Gromov-Wasserstein (srGW) framework and multidimensional scaling by solving the Monge problem in this setting. We also derive novel connections between srGW distance and Gromov-Hausdorff distance. We apply our computational framework to analyze ensembles of political redistricting plans for states with two Congressional districts, achieving an effective visualization of the ensemble as a distribution on a circle which can be used to characterize typical neutral plans, and to flag outliers.",
      "url": "https://www.semanticscholar.org/paper/dc097cdb2866d5bb7776f54e52312b3582a81626",
      "year": 2024,
      "venue": "AAAI Conference on Artificial Intelligence",
      "source": "semantic_scholar",
      "doi": "10.1609/aaai.v39i15.33766",
      "pdf_url": "https://doi.org/10.1609/aaai.v39i15.33766",
      "citations": 6,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 391
    },
    {
      "title": "Control and maintenance of fully-constrained and underconstrained rigid body motion on Lie groups and their tangent bundles",
      "authors": [
        "Brennan S. McCann",
        "Morad Nazari"
      ],
      "abstract": "<p style='text-indent:20px;'>Presented herein are a class of methodologies for conducting constrained motion analysis of rigid bodies within the Udwadia-Kalaba (U-K) formulation. The U-K formulation, primarily devised for systems of particles, is advanced to rigid body dynamics in the geometric mechanics framework and a novel development of U-K formulation for use on nonlinear manifolds, namely the special Euclidean group <inline-formula><tex-math id=\"M1\">\\begin{document}$ {\\mathsf{SE}(3)}$\\end{document}</tex-math></inline-formula> and its second order tangent bundle <inline-formula><tex-math id=\"M2\">\\begin{document}${\\mathsf{T}^2\\mathsf{SE}(3)} $\\end{document}</tex-math></inline-formula>, is proposed in addition to the formulation development on Euclidean spaces. Then, a Morse-Lyapunov based tracking controller using backstepping is applied to capture disturbed initial conditions that the U-K formulation cannot account for. This theoretical development is then applied to fully-constrained and underconstrained scenarios of rigid-body spacecraft motion in a lunar orbit, and the translational and rotational motions of the spacecraft and the control inputs obtained using the proposed methodologies to achieve and maintain those constrained motions are studied.</p>",
      "url": "https://www.semanticscholar.org/paper/0074129b3325b1948ecc3cc72687807e904f6270",
      "year": 2022,
      "venue": "The Journal of Geometric Mechanics",
      "source": "semantic_scholar",
      "doi": "10.3934/jgm.2022002",
      "pdf_url": "https://www.aimsciences.org/article/exportPdf?id=eddeac28-ae6b-4159-952d-d4df118aa9cc",
      "citations": 12,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 392
    },
    {
      "title": "Conjugate Priors and Posterior Inference for the Matrix Langevin Distribution on the Stiefel Manifold",
      "authors": [
        "Subhadip Pal",
        "Subhajit Sengupta",
        "Riten Mitra",
        "Arunava Banerjee"
      ],
      "abstract": "Directional data emerges in a wide array of applications, ranging from atmospheric sciences to medical imaging. Modeling such data, however, poses unique challenges by virtue of their being constrained to non-Euclidean spaces like manifolds. Here, we present a unified Bayesian framework for inference on the Stiefel manifold using the Matrix Langevin distribution. Specifically, we propose a novel family of conjugate priors and establish a number of theoretical properties relevant to statistical inference. Conjugacy enables translation of these properties to their corresponding posteriors, which we exploit to develop the posterior inference scheme. For the implementation of the posterior computation, including the posterior sampling, we adopt a novel computational procedure for evaluating the hypergeometric function of matrix arguments that appears as normalization constants in the relevant densities.",
      "url": "https://www.semanticscholar.org/paper/e4d7be7c65fdd755b006340791547c2f9b298d4c",
      "year": 2020,
      "venue": "",
      "source": "semantic_scholar",
      "doi": "10.1214/19-ba1176",
      "pdf_url": "https://projecteuclid.org/journals/bayesian-analysis/volume-15/issue-3/Conjugate-Priors-and-Posterior-Inference-for-the-Matrix-Langevin-Distribution/10.1214/19-BA1176.pdf",
      "citations": 8,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 393
    },
    {
      "title": "A Unified Gradient-based Framework for Task-agnostic Continual Learning-Unlearning",
      "authors": [
        "Zhehao Huang",
        "Xinwen Cheng",
        "Jie Zhang",
        "Jinghao Zheng",
        "Haoran Wang",
        "Zhengbao He",
        "Tao Li",
        "Xiaolin Huang"
      ],
      "abstract": "Recent advancements in deep models have highlighted the need for intelligent systems that combine continual learning (CL) for knowledge acquisition with machine unlearning (MU) for data removal, forming the Continual Learning-Unlearning (CLU) paradigm. While existing work treats CL and MU as separate processes, we reveal their intrinsic connection through a unified optimization framework based on Kullback-Leibler divergence minimization. This framework decomposes gradient updates for approximate CLU into four components: learning new knowledge, unlearning targeted data, preserving existing knowledge, and modulation via weight saliency. A critical challenge lies in balancing knowledge update and retention during sequential learning-unlearning cycles. To resolve this stability-plasticity dilemma, we introduce a remain-preserved manifold constraint to induce a remaining Hessian compensation for CLU iterations. A fast-slow weight adaptation mechanism is designed to efficiently approximate the second-order optimization direction, combined with adaptive weighting coefficients and a balanced weight saliency mask, proposing a unified implementation framework for gradient-based CLU. Furthermore, we pioneer task-agnostic CLU scenarios that support fine-grained unlearning at the cross-task category and random sample levels beyond the traditional task-aware setups. Experiments demonstrate that the proposed UG-CLU framework effectively coordinates incremental learning, precise unlearning, and knowledge stability across multiple datasets and model architectures, providing a theoretical foundation and methodological support for dynamic, compliant intelligent systems.",
      "url": "https://www.semanticscholar.org/paper/7b46064150ae7a1efed8a6baec5346e1c01d4553",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2505.15178",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 394
    },
    {
      "title": "Optimizing Data Augmentation through Bayesian Model Selection",
      "authors": [
        "Madi Matymov",
        "Ba-Hien Tran",
        "Michael Kampffmeyer",
        "Markus Heinonen",
        "M. Filippone"
      ],
      "abstract": "Data Augmentation (DA) has become an essential tool to improve robustness and generalization of modern machine learning. However, when deciding on DA strategies it is critical to choose parameters carefully, and this can be a daunting task which is traditionally left to trial-and-error or expensive optimization based on validation performance. In this paper, we counter these limitations by proposing a novel framework for optimizing DA. In particular, we take a probabilistic view of DA, which leads to the interpretation of augmentation parameters as model (hyper)-parameters, and the optimization of the marginal likelihood with respect to these parameters as a Bayesian model selection problem. Due to its intractability, we derive a tractable Evidence Lower BOund (ELBO), which allows us to optimize augmentation parameters jointly with model parameters. We provide extensive theoretical results on variational approximation quality, generalization guarantees, invariance properties, and connections to empirical Bayes. Through experiments on computer vision tasks, we show that our approach improves calibration and yields robust performance over fixed or no augmentation. Our work provides a rigorous foundation for optimizing DA through Bayesian principles with significant potential for robust machine learning.",
      "url": "https://www.semanticscholar.org/paper/7407698af584a2073234e6812e215588d3d656c3",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2505.21813",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 395
    },
    {
      "title": "Convex Regularization and Convergence of Policy Gradient Flows under Safety Constraints",
      "authors": [
        "Pekka Malo",
        "L. Viitasaari",
        "Antti-Jussi Suominen",
        "Eeva Vilkkumaa",
        "O. Tahvonen"
      ],
      "abstract": "This paper examines reinforcement learning (RL) in infinite-horizon decision processes with almost-sure safety constraints, crucial for applications like autonomous systems, finance, and resource management. We propose a doubly-regularized RL framework combining reward and parameter regularization to address safety constraints in continuous state-action spaces. The problem is formulated as a convex regularized objective with parametrized policies in the mean-field regime. Leveraging mean-field theory and Wasserstein gradient flows, policies are modeled on an infinite-dimensional statistical manifold, with updates governed by parameter distribution gradient flows. Key contributions include solvability conditions for safety-constrained problems, smooth bounded approximations for gradient flows, and exponential convergence guarantees under sufficient regularization. General regularization conditions, including entropy regularization, support practical particle method implementations. This framework provides robust theoretical insights and guarantees for safe RL in complex, high-dimensional settings.",
      "url": "https://www.semanticscholar.org/paper/7ba834ca2bbfebec6b7da4803b68a731873e991e",
      "year": 2024,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2411.19193",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 396
    },
    {
      "title": "Graph Neural Diffusion via Generalized Opinion Dynamics",
      "authors": [
        "Asela Hevapathige",
        "Asiri Wijesinghe",
        "Ahad N. Zehmakan"
      ],
      "abstract": "There has been a growing interest in developing diffusion-based Graph Neural Networks (GNNs), building on the connections between message passing mechanisms in GNNs and physical diffusion processes. However, existing methods suffer from three critical limitations: (1) they rely on homogeneous diffusion with static dynamics, limiting adaptability to diverse graph structures; (2) their depth is constrained by computational overhead and diminishing interpretability; and (3) theoretical understanding of their convergence behavior remains limited. To address these challenges, we propose GODNF, a Generalized Opinion Dynamics Neural Framework, which unifies multiple opinion dynamics models into a principled, trainable diffusion mechanism. Our framework captures heterogeneous diffusion patterns and temporal dynamics via node-specific behavior modeling and dynamic neighborhood influence, while ensuring efficient and interpretable message propagation even at deep layers. We provide a rigorous theoretical analysis demonstrating GODNF's ability to model diverse convergence configurations. Extensive empirical evaluations of node classification and influence estimation tasks confirm GODNF's superiority over state-of-the-art GNNs.",
      "url": "https://www.semanticscholar.org/paper/546229d0837023e8f183cdc4a3dcded208bd3382",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2508.11249",
      "pdf_url": "",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 397
    },
    {
      "title": "cryoSENSE: Compressive Sensing Enables High-throughput Microscopy with Sparse and Generative Priors on the Protein Cryo-EM Image Manifold",
      "authors": [
        "Zain Shabeeb",
        "Daniel Saeedi",
        "Darin Tsui",
        "Vida Jamali",
        "Amirali Aghazadeh"
      ],
      "abstract": "Cryo-electron microscopy (cryo-EM) enables the atomic-resolution visualization of biomolecules; however, modern direct detectors generate data volumes that far exceed the available storage and transfer bandwidth, thereby constraining practical throughput. We introduce cryoSENSE, the computational realization of a hardware-software co-designed framework for compressive cryo-EM sensing and acquisition. We show that cryo-EM images of proteins lie on low-dimensional manifolds that can be independently represented using sparse priors in predefined bases and generative priors captured by a denoising diffusion model. cryoSENSE leverages these low-dimensional manifolds to enable faithful image reconstruction from spatial and Fourier-domain undersampled measurements while preserving downstream structural resolution. In experiments, cryoSENSE increases acquisition throughput by up to 2.5$\\times$ while retaining the original 3D resolution, offering controllable trade-offs between the number of masked measurements and the level of downsampling. Sparse priors favor faithful reconstruction from Fourier-domain measurements and moderate compression, whereas generative diffusion priors achieve accurate recovery from pixel-domain measurements and more severe undersampling. Project website: https://cryosense.github.io.",
      "url": "http://arxiv.org/abs/2511.12931v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.12931v2",
      "citations": null,
      "categories": [
        "eess.IV",
        "q-bio.BM"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 398
    },
    {
      "title": "Channel-Constrained Markovian Quantum Diffusion Model from Open System Perspective",
      "authors": [
        "Qin-Sheng Zhu",
        "Geng Chen",
        "Lian-Hui Yu",
        "Xiaodong Xing",
        "Xiao-Yu Li"
      ],
      "abstract": "We present a channel-constrained Markovian quantum diffusion (CCMQD) model that prepares quantum states by rigorously framing the generative process within the dynamics of open quantum systems. Our model interprets the forward diffusion process as natural decoherence using quantum master equations, whereas the reverse denoising is achieved by learning inverse quantum channels. Our core innovation is a comprehensive channel-constrained framework: we model the diffusion and denoising steps as quantum channels defined by Kraus operators, ensure their physical validity through optimization on the Stiefel manifold, and introduce tailored training strategies and loss functions that leverage this constrained structure for high-fidelity state reconstruction. Experimental validation on systems ranging from single qubits to entangled states $7$ -qubits demonstrates high-fidelity state generation, achieving fidelities exceeding $0.998$ under both random and depolarizing noise conditions. This work confirms that quantum diffusion can be characterized as a controlled Markov evolution, demonstrating that environmental interactions are not limited to being a source of decoherence but can also be utilized to achieve high-fidelity quantum state synthesis.",
      "url": "http://arxiv.org/abs/2511.12221v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.12221v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 399
    },
    {
      "title": "Formation of Close Binaries through Massive Black Hole Perturbations and Chaotic Tides",
      "authors": [
        "Howard Hao-Tse Huang",
        "Wenbin Lu"
      ],
      "abstract": "Hills breakup of binary systems allows massive black holes (MBH) to produce hyper-velocity stars (HVSs) and tightly bound stars. The long timescale of orbital relaxation means that binaries must spend numerous orbits around the MBH before they are tidally broken apart. Repeated MBH tidal perturbations over multiple pericenter passages can perturb the binary inner orbit to high eccentricities, leading to strong tidal interactions between the stars. In this work, we develop a physical model of the MBH-binary system, taking into account outer orbital relaxation, MBH tidal perturbations, and tidal interactions between the binaries in the form of dynamical tides. We show that when the inner orbit reaches high eccentricities such that the pericenter radius is only a few times stellar radii ($R_*$), the stellar oscillation modes can grow chaotically and rapidly harden the binaries to semi-major axes $a_b\\lesssim 10\\,R_*$. We find that a significant fraction (up to 50\\%) of initially wide binaries that are in the empty loss-cone regime ($a_b\\sim 1.0\\,{\\rm AU}$) do not undergo Hills breakup as wide binaries, but instead experience chaotic growth of tides and become close binaries. These tidally hardened binaries provide a new channel for the production of the fastest HVSs, and are connected to other nuclear transients such as repeating partial tidal disruption events and quasi-periodic eruptions.",
      "url": "http://arxiv.org/abs/2511.11965v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.11965v1",
      "citations": null,
      "categories": [
        "astro-ph.GA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 400
    },
    {
      "title": "Reaching for the Edge II: Stellar Halos out to Large Radii as a Tracer of Dark Matter Halo Mass",
      "authors": [
        "Katya Leidig",
        "Benedikt Diemer",
        "Song Huang",
        "Shuo Xu",
        "Conghao Zhou",
        "Alexie Leauthaud"
      ],
      "abstract": "The diffuse outskirts of brightest cluster galaxies (BCGs) encode valuable information about the assembly history and mass of their host dark matter halos. However, the low surface brightness of these stellar halos has historically made them difficult to observe. Recent deep imaging, particularly with Hyper Suprime-Cam (HSC), has shown that the stellar mass within relatively large projected annuli, such as within $50$ and $100$ kpc, is a promising proxy for halo mass. However, the optimal radial definition of this \"outskirt mass\" remains uncertain. We construct an HSC-like mock observing pipeline to measure the stellar mass density profiles of BCGs in the IllustrisTNG simulations. Our mock observations closely reproduce HSC profiles across six orders of magnitude in surface density. We then systematically measure stellar masses within different annuli and how tightly they are connected to halo mass. We find that stellar masses measured within simple apertures exhibit considerably more scatter in the stellar mass-halo mass relation than those measured within projected ellipsoidal annuli. We identify an optimal range of definitions, with inner radii between $\\sim 70$-$200$ kpc and outer radii between $\\sim 125$-$500$ kpc. We also introduce two halo-mass-dependent Sérsic models for the average stellar halo profiles. We present a Sérsic-based fitting function that describes the profiles as a function of the halo mass, $M_{\\rm vir}$, with a median error of $54\\%$. Adding the central stellar mass of the BCG as a second parameter slightly improves the accuracy to a median error of $39\\%$. Together, these results provide fitting functions for BCG stellar halos that can be applied to future wide-field surveys to infer halo masses from deep imaging data.",
      "url": "http://arxiv.org/abs/2511.10723v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.10723v1",
      "citations": null,
      "categories": [
        "astro-ph.GA",
        "astro-ph.CO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 401
    },
    {
      "title": "Deformations of Locally Conformal Spin(7) Instantons",
      "authors": [
        "Eyup Yalcinkaya"
      ],
      "abstract": "We explore the deformation theory of instantons on locally conformal (LC) $Spin(7)$ manifolds. These structures, characterized by a non-parallel fundamental 4-form $Φ$ satisfying $dΦ= θ\\wedge Φ$, represent a significant, yet geometrically constrained, class of non-integrable $G$-structures. We analyze the infinitesimal deformation complex for $Spin(7)$-instantons in this setting.\n  Our primary contribution is the reformulation of the linearized deformation equations -- comprising the linearized instanton condition and a gauge-fixing term -- using a $t$-parameter family of Dirac operators. We demonstrate that the $t$-dependent torsion terms arising from the Lee form $θ$ cancel precisely. This unexpected simplification reveals that the deformation space $\\mathcal{H}^1$ is governed entirely by the Levi-Civita geometry, effectively reducing the torsion-full problem to a more classical, torsion-free (Levi-Civita) setting.\n  Using a Lichnerowicz-type rigidity theorem, we establish a general condition for an (LC) $Spin(7)$-instanton to be rigid (i.e., $\\mathcal{H}^1 = \\{0\\}$). We apply this theory to the flat instanton ($A=0$) on known compact homogeneous (LC) $Spin(7)$ manifolds and conclude that the flat instanton on these spaces is non-rigid, thus possessing a non-trivial moduli space.",
      "url": "http://arxiv.org/abs/2511.09161v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.09161v1",
      "citations": null,
      "categories": [
        "math.DG",
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 402
    },
    {
      "title": "Tube Integrability in a Time-Dependent Nonlinear Oscillator",
      "authors": [
        "Johannes Hagel"
      ],
      "abstract": "We study the nonlinear oscillator z'' + omega^2 z + g(t) z^2 = 0 with a time-dependent coefficient g(t). We show that this equation admits an exact quadratic invariant I(z,p,t) provided that g(t) = alpha2(t)^(-5/2) and that alpha2(t) satisfies a nonlinear third-order differential equation. The resulting invariant constrains the dynamics to a smooth two-dimensional surface in the extended phase space (z,p,t). If alpha2(t) is periodic, this surface forms a compact invariant torus. However, we show that periodic solutions of alpha2(t) are generically obstructed by a resonance mechanism, leading instead to an aperiodic but non-chaotic evolution. In this regime the invariant surface is non-compact and extends along the time direction, forming a tube rather than a torus. We therefore propose the term \"tube integrability\" for integrable systems whose invariant manifolds are non-compact in time. A perturbation expansion for alpha2(t) up to third order is derived and compared with numerical integration, clarifying the parameter regimes in which the truncated series provides quantitatively accurate approximations. The breakdown of the series for small y0 reflects the asymptotic nature of the expansion rather than a loss of integrability.",
      "url": "http://arxiv.org/abs/2511.13740v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.13740v1",
      "citations": null,
      "categories": [
        "math.DS",
        "nlin.SI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 403
    },
    {
      "title": "Contact Wasserstein Geodesics for Non-Conservative Schrödinger Bridges",
      "authors": [
        "Andrea Testa",
        "Søren Hauberg",
        "Tamim Asfour",
        "Leonel Rozo"
      ],
      "abstract": "The Schrödinger Bridge provides a principled framework for modeling stochastic processes between distributions; however, existing methods are limited by energy-conservation assumptions, which constrains the bridge's shape preventing it from model varying-energy phenomena. To overcome this, we introduce the non-conservative generalized Schrödinger bridge (NCGSB), a novel, energy-varying reformulation based on contact Hamiltonian mechanics. By allowing energy to change over time, the NCGSB provides a broader class of real-world stochastic processes, capturing richer and more faithful intermediate dynamics. By parameterizing the Wasserstein manifold, we lift the bridge problem to a tractable geodesic computation in a finite-dimensional space. Unlike computationally expensive iterative solutions, our contact Wasserstein geodesic (CWG) is naturally implemented via a ResNet architecture and relies on a non-iterative solver with near-linear complexity. Furthermore, CWG supports guided generation by modulating a task-specific distance metric. We validate our framework on tasks including manifold navigation, molecular dynamics predictions, and image generation, demonstrating its practical benefits and versatility.",
      "url": "http://arxiv.org/abs/2511.06856v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.06856v2",
      "citations": null,
      "categories": [
        "cs.LG",
        "math.DG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 404
    },
    {
      "title": "Non-Negative Stiefel Approximating Flow: Orthogonalish Matrix Optimization for Interpretable Embeddings",
      "authors": [
        "Brian B. Avants",
        "Nicholas J. Tustison",
        "James R Stone"
      ],
      "abstract": "Interpretable representation learning is a central challenge in modern machine learning, particularly in high-dimensional settings such as neuroimaging, genomics, and text analysis. Current methods often struggle to balance the competing demands of interpretability and model flexibility, limiting their effectiveness in extracting meaningful insights from complex data. We introduce Non-negative Stiefel Approximating Flow (NSA-Flow), a general-purpose matrix estimation framework that unifies ideas from sparse matrix factorization, orthogonalization, and constrained manifold learning. NSA-Flow enforces structured sparsity through a continuous balance between reconstruction fidelity and column-wise decorrelation, parameterized by a single tunable weight. The method operates as a smooth flow near the Stiefel manifold with proximal updates for non-negativity and adaptive gradient control, yielding representations that are simultaneously sparse, stable, and interpretable. Unlike classical regularization schemes, NSA-Flow provides an intuitive geometric mechanism for manipulating sparsity at the level of global structure while simplifying latent features. We demonstrate that the NSA-Flow objective can be optimized smoothly and integrates seamlessly with existing pipelines for dimensionality reduction while improving interpretability and generalization in both simulated and real biomedical data. Empirical validation on the Golub leukemia dataset and in Alzheimer's disease demonstrate that the NSA-Flow constraints can maintain or improve performance over related methods with little additional methodological effort. NSA-Flow offers a scalable, general-purpose tool for interpretable ML, applicable across data science domains.",
      "url": "http://arxiv.org/abs/2511.06425v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.06425v1",
      "citations": null,
      "categories": [
        "stat.ML",
        "cs.CV",
        "cs.LG",
        "stat.ME"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 405
    },
    {
      "title": "Geometrically robust least squares through manifold optimization",
      "authors": [
        "Jeremy Coulson",
        "Alberto Padoan",
        "Cyrus Mostajeran"
      ],
      "abstract": "This paper presents a methodology for solving a geometrically robust least squares problem, which arises in various applications where the model is subject to geometric constraints. The problem is formulated as a minimax optimization problem on a product manifold, where one variable is constrained to a ball describing uncertainty. To handle the constraint, an exact penalty method is applied. A first-order gradient descent ascent algorithm is proposed to solve the problem, and its convergence properties are illustrated by an example. The proposed method offers a robust approach to solving a wide range of problems arising in signal processing and data-driven control.",
      "url": "http://arxiv.org/abs/2511.03644v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.03644v1",
      "citations": null,
      "categories": [
        "math.OC",
        "eess.SY"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 406
    },
    {
      "title": "Manifold-constrained Hamilton-Jacobi Reachability Learning for Decentralized Multi-Agent Motion Planning",
      "authors": [
        "Qingyi Chen",
        "Ruiqi Ni",
        "Jun Kim",
        "Ahmed H. Qureshi"
      ],
      "abstract": "Safe multi-agent motion planning (MAMP) under task-induced constraints is a critical challenge in robotics. Many real-world scenarios require robots to navigate dynamic environments while adhering to manifold constraints imposed by tasks. For example, service robots must carry cups upright while avoiding collisions with humans or other robots. Despite recent advances in decentralized MAMP for high-dimensional systems, incorporating manifold constraints remains difficult. To address this, we propose a manifold-constrained Hamilton-Jacobi reachability (HJR) learning framework for decentralized MAMP. Our method solves HJR problems under manifold constraints to capture task-aware safety conditions, which are then integrated into a decentralized trajectory optimization planner. This enables robots to generate motion plans that are both safe and task-feasible without requiring assumptions about other agents' policies. Our approach generalizes across diverse manifold-constrained tasks and scales effectively to high-dimensional multi-agent manipulation problems. Experiments show that our method outperforms existing constrained motion planners and operates at speeds suitable for real-world applications. Video demonstrations are available at https://youtu.be/RYcEHMnPTH8 .",
      "url": "http://arxiv.org/abs/2511.03591v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.03591v1",
      "citations": null,
      "categories": [
        "cs.RO",
        "eess.SY"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 407
    },
    {
      "title": "SKGE: Spherical Knowledge Graph Embedding with Geometric Regularization",
      "authors": [
        "Xuan-Truong Quan",
        "Xuan-Son Quan",
        "Duc Do Minh",
        "Vinh Nguyen Van"
      ],
      "abstract": "Knowledge graph embedding (KGE) has become a fundamental technique for representation learning on multi-relational data. Many seminal models, such as TransE, operate in an unbounded Euclidean space, which presents inherent limitations in modeling complex relations and can lead to inefficient training. In this paper, we propose Spherical Knowledge Graph Embedding (SKGE), a model that challenges this paradigm by constraining entity representations to a compact manifold: a hypersphere. SKGE employs a learnable, non-linear Spherization Layer to map entities onto the sphere and interprets relations as a hybrid translate-then-project transformation. Through extensive experiments on three benchmark datasets, FB15k-237, CoDEx-S, and CoDEx-M, we demonstrate that SKGE consistently and significantly outperforms its strong Euclidean counterpart, TransE, particularly on large-scale benchmarks such as FB15k-237 and CoDEx-M, demonstrating the efficacy of the spherical geometric prior. We provide an in-depth analysis to reveal the sources of this advantage, showing that this geometric constraint acts as a powerful regularizer, leading to comprehensive performance gains across all relation types. More fundamentally, we prove that the spherical geometry creates an \"inherently hard negative sampling\" environment, naturally eliminating trivial negatives and forcing the model to learn more robust and semantically coherent representations. Our findings compellingly demonstrate that the choice of manifold is not merely an implementation detail but a fundamental design principle, advocating for geometric priors as a cornerstone for designing the next generation of powerful and stable KGE models.",
      "url": "http://arxiv.org/abs/2511.02460v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.02460v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 408
    },
    {
      "title": "SE(3)-PoseFlow: Estimating 6D Pose Distributions for Uncertainty-Aware Robotic Manipulation",
      "authors": [
        "Yufeng Jin",
        "Niklas Funk",
        "Vignesh Prasad",
        "Zechu Li",
        "Mathias Franzius",
        "Jan Peters",
        "Georgia Chalvatzaki"
      ],
      "abstract": "Object pose estimation is a fundamental problem in robotics and computer vision, yet it remains challenging due to partial observability, occlusions, and object symmetries, which inevitably lead to pose ambiguity and multiple hypotheses consistent with the same observation. While deterministic deep networks achieve impressive performance under well-constrained conditions, they are often overconfident and fail to capture the multi-modality of the underlying pose distribution. To address these challenges, we propose a novel probabilistic framework that leverages flow matching on the SE(3) manifold for estimating 6D object pose distributions. Unlike existing methods that regress a single deterministic output, our approach models the full pose distribution with a sample-based estimate and enables reasoning about uncertainty in ambiguous cases such as symmetric objects or severe occlusions. We achieve state-of-the-art results on Real275, YCB-V, and LM-O, and demonstrate how our sample-based pose estimates can be leveraged in downstream robotic manipulation tasks such as active perception for disambiguating uncertain viewpoints or guiding grasp synthesis in an uncertainty-aware manner.",
      "url": "http://arxiv.org/abs/2511.01501v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.01501v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.RO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 409
    },
    {
      "title": "The Geometry of Grokking: Norm Minimization on the Zero-Loss Manifold",
      "authors": [
        "Tiberiu Musat"
      ],
      "abstract": "Grokking is a puzzling phenomenon in neural networks where full generalization occurs only after a substantial delay following the complete memorization of the training data. Previous research has linked this delayed generalization to representation learning driven by weight decay, but the precise underlying dynamics remain elusive. In this paper, we argue that post-memorization learning can be understood through the lens of constrained optimization: gradient descent effectively minimizes the weight norm on the zero-loss manifold. We formally prove this in the limit of infinitesimally small learning rates and weight decay coefficients. To further dissect this regime, we introduce an approximation that decouples the learning dynamics of a subset of parameters from the rest of the network. Applying this framework, we derive a closed-form expression for the post-memorization dynamics of the first layer in a two-layer network. Experiments confirm that simulating the training process using our predicted gradients reproduces both the delayed generalization and representation learning characteristic of grokking.",
      "url": "http://arxiv.org/abs/2511.01938v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2511.01938v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 410
    },
    {
      "title": "Robust Graph Condensation via Classification Complexity Mitigation",
      "authors": [
        "Jiayi Luo",
        "Qingyun Sun",
        "Beining Yang",
        "Haonan Yuan",
        "Xingcheng Fu",
        "Yanbiao Ma",
        "Jianxin Li",
        "Philip S. Yu"
      ],
      "abstract": "Graph condensation (GC) has gained significant attention for its ability to synthesize smaller yet informative graphs. However, existing studies often overlook the robustness of GC in scenarios where the original graph is corrupted. In such cases, we observe that the performance of GC deteriorates significantly, while existing robust graph learning technologies offer only limited effectiveness. Through both empirical investigation and theoretical analysis, we reveal that GC is inherently an intrinsic-dimension-reducing process, synthesizing a condensed graph with lower classification complexity. Although this property is critical for effective GC performance, it remains highly vulnerable to adversarial perturbations. To tackle this vulnerability and improve GC robustness, we adopt the geometry perspective of graph data manifold and propose a novel Manifold-constrained Robust Graph Condensation framework named MRGC. Specifically, we introduce three graph data manifold learning modules that guide the condensed graph to lie within a smooth, low-dimensional manifold with minimal class ambiguity, thereby preserving the classification complexity reduction capability of GC and ensuring robust performance under universal adversarial attacks. Extensive experiments demonstrate the robustness of \\ModelName\\ across diverse attack scenarios.",
      "url": "http://arxiv.org/abs/2510.26451v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.26451v2",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 411
    },
    {
      "title": "Causal-Aware Generative Adversarial Networks with Reinforcement Learning",
      "authors": [
        "Tu Anh Hoang Nguyen",
        "Dang Nguyen",
        "Tri-Nhan Vo",
        "Thuc Duy Le",
        "Sunil Gupta"
      ],
      "abstract": "The utility of tabular data for tasks ranging from model training to large-scale data analysis is often constrained by privacy concerns or regulatory hurdles. While existing data generation methods, particularly those based on Generative Adversarial Networks (GANs), have shown promise, they frequently struggle with capturing complex causal relationship, maintaining data utility, and providing provable privacy guarantees suitable for enterprise deployment. We introduce CA-GAN, a novel generative framework specifically engineered to address these challenges for real-world tabular datasets. CA-GAN utilizes a two-step approach: causal graph extraction to learn a robust, comprehensive causal relationship in the data's manifold, followed by a custom Conditional WGAN-GP (Wasserstein GAN with Gradient Penalty) that operates exclusively as per the structure of nodes in the causal graph. More importantly, the generator is trained with a new Reinforcement Learning-based objective that aligns the causal graphs constructed from real and fake data, ensuring the causal awareness in both training and sampling phases. We demonstrate CA-GAN superiority over six SOTA methods across 14 tabular datasets. Our evaluations, focused on core data engineering metrics: causal preservation, utility preservation, and privacy preservation. Our method offers a practical, high-performance solution for data engineers seeking to create high-quality, privacy-compliant synthetic datasets to benchmark database systems, accelerate software development, and facilitate secure data-driven research.",
      "url": "http://arxiv.org/abs/2510.24046v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.24046v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 412
    },
    {
      "title": "The Gravitational Aspect of Information: The Physical Reality of Asymmetric \"Distance\"",
      "authors": [
        "Tomoi Koide",
        "Armin van de Venn"
      ],
      "abstract": "We show that when a Brownian bridge is physically constrained to satisfy a canonical condition, its time evolution exactly coincides with an m-geodesic on the statistical manifold of Gaussian distributions. This identification provides a direct physical realization of a geometric concept in information geometry. It implies that purely random processes evolve along informationally straight trajectories, analogous to geodesics in general relativity. Our findings suggest that the asymmetry of informational ``distance\" (divergence) plays a fundamental physical role, offering a concrete step toward an equivalence principle for information.",
      "url": "http://arxiv.org/abs/2510.22664v3",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.22664v3",
      "citations": null,
      "categories": [
        "cond-mat.stat-mech",
        "cs.IT",
        "gr-qc",
        "hep-ph",
        "math.ST",
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 413
    },
    {
      "title": "If You Want to Be Robust, Be Wary of Initialization",
      "authors": [
        "Sofiane Ennadir",
        "Johannes F. Lutzeyer",
        "Michalis Vazirgiannis",
        "El Houcine Bergou"
      ],
      "abstract": "Graph Neural Networks (GNNs) have demonstrated remarkable performance across a spectrum of graph-related tasks, however concerns persist regarding their vulnerability to adversarial perturbations. While prevailing defense strategies focus primarily on pre-processing techniques and adaptive message-passing schemes, this study delves into an under-explored dimension: the impact of weight initialization and associated hyper-parameters, such as training epochs, on a model's robustness. We introduce a theoretical framework bridging the connection between initialization strategies and a network's resilience to adversarial perturbations. Our analysis reveals a direct relationship between initial weights, number of training epochs and the model's vulnerability, offering new insights into adversarial robustness beyond conventional defense mechanisms. While our primary focus is on GNNs, we extend our theoretical framework, providing a general upper-bound applicable to Deep Neural Networks. Extensive experiments, spanning diverse models and real-world datasets subjected to various adversarial attacks, validate our findings. We illustrate that selecting appropriate initialization not only ensures performance on clean datasets but also enhances model robustness against adversarial perturbations, with observed gaps of up to 50\\% compared to alternative initialization approaches.",
      "url": "http://arxiv.org/abs/2510.22652v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.22652v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 414
    },
    {
      "title": "Confidence is Not Competence",
      "authors": [
        "Debdeep Sanyal",
        "Manya Pandey",
        "Dhruv Kumar",
        "Saurabh Deshpande",
        "Murari Mandal"
      ],
      "abstract": "Large language models (LLMs) often exhibit a puzzling disconnect between their asserted confidence and actual problem-solving competence. We offer a mechanistic account of this decoupling by analyzing the geometry of internal states across two phases - pre-generative assessment and solution execution. A simple linear probe decodes the internal \"solvability belief\" of a model, revealing a well-ordered belief axis that generalizes across model families and across math, code, planning, and logic tasks. Yet, the geometries diverge - although belief is linearly decodable, the assessment manifold has high linear effective dimensionality as measured from the principal components, while the subsequent reasoning trace evolves on a much lower-dimensional manifold. This sharp reduction in geometric complexity from thought to action mechanistically explains the confidence-competence gap. Causal interventions that steer representations along the belief axis leave final solutions unchanged, indicating that linear nudges in the complex assessment space do not control the constrained dynamics of execution. We thus uncover a two-system architecture - a geometrically complex assessor feeding a geometrically simple executor. These results challenge the assumption that decodable beliefs are actionable levers, instead arguing for interventions that target the procedural dynamics of execution rather than the high-level geometry of assessment.",
      "url": "http://arxiv.org/abs/2510.24772v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.24772v1",
      "citations": null,
      "categories": [
        "cs.CL",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 415
    },
    {
      "title": "Landau Polarons as Generators of Quantum-Coherent States",
      "authors": [
        "Arnab Ghosh",
        "Patrick Brosseau",
        "Dmitry N. Dirin",
        "Rui Tao",
        "Maksym V. Kovalenko",
        "Patanjali Kambhampati"
      ],
      "abstract": "Since Landau's theory, polarons have been understood as quasiparticles in which charges are dressed by the lattice field, yet decades of transport and spectroscopic studies have yielded only static indirect renormalizations. Whether such dressing can dynamically reorganize electronic spectra to generate new quantum-coherent states has remained unresolved. Here we use femtosecond coherent multidimensional spectroscopy on size and composition controlled perovskite quantum dots to track polaronic field-induced dynamics in real time, revealing their consequences. We observe a delayed condensation into a confined spectrum of coherent states on 50-150 fs timescales, with couplings between these states evolving dynamically on the same timescale. The splittings are robust, exhibit anomalous linear size dependence, exceed single-particle splittings and manifest at 300 K. A Raman-constrained spin-boson Hamiltonian captures both the anomalous scaling and dynamical onset, establishing polarons as generators of coherent manifolds that enable collective quantum phenomena including superradiance, superfluorescence and superabsorption.",
      "url": "http://arxiv.org/abs/2510.20962v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.20962v1",
      "citations": null,
      "categories": [
        "cond-mat.mes-hall",
        "cond-mat.mtrl-sci",
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 416
    },
    {
      "title": "Bayesian Prediction under Moment Conditioning",
      "authors": [
        "Nicholas G. Polson",
        "Daniel Zantedeschi"
      ],
      "abstract": "Prediction is a central task of statistics and machine learning, yet many inferential settings provide only partial information, typically in the form of moment constraints or estimating equations. We develop a finite, fully Bayesian framework for propagating such partial information through predictive distributions. Building on de Finetti's representation theorem, we construct a curvature-adaptive version of exchangeable updating that operates directly under finite constraints, yielding an explicit discrete-Gaussian mixture that quantifies predictive uncertainty. The resulting finite-sample bounds depend on the smallest eigenvalue of the information-geometric Hessian, which measures the curvature and identification strength of the constraint manifold. This approach unifies empirical likelihood, Bayesian empirical likelihood, and generalized method-of-moments estimation within a common predictive geometry. On the operational side, it provides computable curvature-sensitive uncertainty bounds for constrained prediction; on the theoretical side, it recovers de Finetti's coherence, Doob's martingale convergence and local asymptotic normality as limiting cases of the same finite mechanism. Our framework thus offers a constructive bridge between partial information and full Bayesian prediction.",
      "url": "http://arxiv.org/abs/2510.20742v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.20742v1",
      "citations": null,
      "categories": [
        "math.ST"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 417
    },
    {
      "title": "PDE-Free Mass-Constrained Learning of Complex Systems with Hidden States: The crowd dynamics case",
      "authors": [
        "Gianmaria Viola",
        "Alessandro Della Pia",
        "Lucia Russo",
        "Ioannis Kevrekidis",
        "Constantinos Siettos"
      ],
      "abstract": "We propose a machine learning framework based on the next-generation Equation-Free algorithm for learning the spatio-temporal dynamics of mass-constrained complex systems with hidden states, whose dynamics can in principle be described by PDEs, but lack explicit models. In these cases, some variables, closures, and potentials governing the dynamics are generally not directly observable and therefore must be inferred from data. Here, we construct manifold-ROMs -- using delayed coordinates, thus exploiting the Takens'/Whitney's embedding theorems. In the first stage, we employ both linear (POD) and nonlinear manifold learning (Diffusion Maps, DMs) to extract low-dimensional latent representations of the complex spatio-temporal evolution. In the second step, we learn predictive manifold-informed ROMs to approximate the solution operator on the latent space. In the final step, the latent dynamics are lifted back to the original high-dimensional space by solving a pre-image problem. We prove that both POD and the particular $k$-nearest neighbors lifting operators preserve the mass, a crucial property in the context of many problems, including computational fluid dynamics (CFD) and crowd dynamics. Actually, the proposed framework reconstructs the solution operator of the unavailable mass-constrained PDE, bypassing the need to discover an explicit form of the PDE per se. We demonstrate our approach via the Hughes model, approximating the dynamics of individuals minimizing travel time while avoiding obstacles and high-density regions. We show that DMs-informed ROMs outperform the best POD-informed ROMs thus resulting in stable and accurate approximations of the solution operator both in the latent space and, via reconstruction, in the high-dimensional space, and can therefore be integrated reliably over long time horizons.",
      "url": "http://arxiv.org/abs/2510.17657v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.17657v2",
      "citations": null,
      "categories": [
        "math.NA",
        "physics.flu-dyn"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 418
    },
    {
      "title": "3DPR: Single Image 3D Portrait Relight using Generative Priors",
      "authors": [
        "Pramod Rao",
        "Abhimitra Meka",
        "Xilong Zhou",
        "Gereon Fox",
        "Mallikarjun B R",
        "Fangneng Zhan",
        "Tim Weyrich",
        "Bernd Bickel",
        "Hanspeter Pfister",
        "Wojciech Matusik"
      ],
      "abstract": "Rendering novel, relit views of a human head, given a monocular portrait image as input, is an inherently underconstrained problem. The traditional graphics solution is to explicitly decompose the input image into geometry, material and lighting via differentiable rendering; but this is constrained by the multiple assumptions and approximations of the underlying models and parameterizations of these scene components. We propose 3DPR, an image-based relighting model that leverages generative priors learnt from multi-view One-Light-at-A-Time (OLAT) images captured in a light stage. We introduce a new diverse and large-scale multi-view 4K OLAT dataset of 139 subjects to learn a high-quality prior over the distribution of high-frequency face reflectance. We leverage the latent space of a pre-trained generative head model that provides a rich prior over face geometry learnt from in-the-wild image datasets. The input portrait is first embedded in the latent manifold of such a model through an encoder-based inversion process. Then a novel triplane-based reflectance network trained on our lightstage data is used to synthesize high-fidelity OLAT images to enable image-based relighting. Our reflectance network operates in the latent space of the generative head model, crucially enabling a relatively small number of lightstage images to train the reflectance model. Combining the generated OLATs according to a given HDRI environment maps yields physically accurate environmental relighting results. Through quantitative and qualitative evaluations, we demonstrate that 3DPR outperforms previous methods, particularly in preserving identity and in capturing lighting effects such as specularities, self-shadows, and subsurface scattering. Project Page: https://vcai.mpi-inf.mpg.de/projects/3dpr/",
      "url": "http://arxiv.org/abs/2510.15846v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.15846v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 419
    },
    {
      "title": "Hypergame-based Cognition Modeling and Intention Interpretation for Human-Driven Vehicles in Connected Mixed Traffic",
      "authors": [
        "Jianguo Chen",
        "Zhengqin Liu",
        "Jinlong Lei",
        "Peng Yi",
        "Yiguang Hong",
        "Hong Chen"
      ],
      "abstract": "With the practical implementation of connected and autonomous vehicles (CAVs), the traffic system is expected to remain a mix of CAVs and human-driven vehicles (HVs) for the foreseeable future. To enhance safety and traffic efficiency, the trajectory planning strategies of CAVs must account for the influence of HVs, necessitating accurate HV trajectory prediction. Current research often assumes that human drivers have perfect knowledge of all vehicles' objectives, an unrealistic premise. This paper bridges the gap by leveraging hypergame theory to account for cognitive and perception limitations in HVs. We model human bounded rationality without assuming them to be merely passive followers and propose a hierarchical cognition modeling framework that captures cognitive relationships among vehicles. We further analyze the cognitive stability of the system, proving that the strategy profile where all vehicles adopt cognitively equilibrium strategies constitutes a hyper Nash equilibrium when CAVs accurately learn HV parameters. To achieve this, we develop an inverse learning algorithm for distributed intention interpretation via vehicle-to-everything (V2X) communication, which extends the framework to both offline and online scenarios. Additionally, we introduce a distributed trajectory prediction and planning approach for CAVs, leveraging the learned parameters in real time. Simulations in highway lane-changing scenarios demonstrate the proposed method's accuracy in parameter learning, robustness to noisy trajectory observations, and safety in HV trajectory prediction. The results validate the effectiveness of our method in both offline and online implementations.",
      "url": "http://arxiv.org/abs/2510.15573v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.15573v1",
      "citations": null,
      "categories": [
        "eess.SY",
        "cs.MA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 420
    },
    {
      "title": "Riemannian Bilevel Optimization with Gradient Aggregation",
      "authors": [
        "Zhuo Chen",
        "Xinjian Xu",
        "Shihui Ying",
        "Tieyong Zeng"
      ],
      "abstract": "Bilevel optimization (BLO) offers a principled framework for hierarchical decision-making and has been widely applied in machine learning tasks such as hyperparameter optimization and meta-learning. While existing BLO methods are mostly developed in Euclidean spaces, many real-world problems involve structural constraints. In this paper, we propose a Riemannian bilevel optimization (RBLO) algorithm that incorporates a bilevel descent aggregation (BDA) scheme to jointly coordinate upper- and lower-level updates. Concretely, first we abstract the constraints in the BLO to a manifold structure and then transform the constrained BLO be a unconstrained RBLO problem. Second, to address limitations of existing RBLO methods, particularly the restrictive assumptions required for convergence, we reformulate the bilevel problem using smooth manifold mappings and provide a convergence analysis under the conditions of geodesic convexity and Lipschitz smoothness. Finally, we recall the multi-view hypergraph spectral clustering task, and evaluate the proposed approach on 3sources data sets. The numerical results validate the superior performance over Euclidean and manifold-based baselines.",
      "url": "http://arxiv.org/abs/2510.15305v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.15305v1",
      "citations": null,
      "categories": [
        "math.OC",
        "math.NA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 421
    },
    {
      "title": "Learning an Image Editing Model without Image Editing Pairs",
      "authors": [
        "Nupur Kumari",
        "Sheng-Yu Wang",
        "Nanxuan Zhao",
        "Yotam Nitzan",
        "Yuheng Li",
        "Krishna Kumar Singh",
        "Richard Zhang",
        "Eli Shechtman",
        "Jun-Yan Zhu",
        "Xun Huang"
      ],
      "abstract": "Recent image editing models have achieved impressive results while following natural language editing instructions, but they rely on supervised fine-tuning with large datasets of input-target pairs. This is a critical bottleneck, as such naturally occurring pairs are hard to curate at scale. Current workarounds use synthetic training pairs that leverage the zero-shot capabilities of existing models. However, this can propagate and magnify the artifacts of the pretrained model into the final trained model. In this work, we present a new training paradigm that eliminates the need for paired data entirely. Our approach directly optimizes a few-step diffusion model by unrolling it during training and leveraging feedback from vision-language models (VLMs). For each input and editing instruction, the VLM evaluates if an edit follows the instruction and preserves unchanged content, providing direct gradients for end-to-end optimization. To ensure visual fidelity, we incorporate distribution matching loss (DMD), which constrains generated images to remain within the image manifold learned by pretrained models. We evaluate our method on standard benchmarks and include an extensive ablation study. Without any paired data, our method performs on par with various image editing diffusion models trained on extensive supervised paired data, under the few-step setting. Given the same VLM as the reward model, we also outperform RL-based techniques like Flow-GRPO.",
      "url": "http://arxiv.org/abs/2510.14978v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.14978v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 422
    },
    {
      "title": "X-ray panorama of the SS433/W50 complex by SRG/eROSITA",
      "authors": [
        "Rashid Sunyaev",
        "Ildar Khabibullin",
        "Eugene Churazov",
        "Marat Gilfanov",
        "Pavel Medvedev",
        "Sergey Sazonov"
      ],
      "abstract": "Galactic microquasar SS433 and the radio nebula W50 surrounding it present a prototypical example of a hyper-Eddington binary system shaping its ambient interstellar medium via energetic outflows. In this paper, we present X-ray observations of the SS433/W50 complex by the eROSITA telescope onboard the SRG space observatory. These data provide images of the entire nebula characterized by a very large dynamic range and allow spectral analysis of the diffuse X-ray emission. In particular, these data illustrate a close connection between the thermal and non-thermal components of W50 on scales ranging from sub-parsecs, represented by narrow X-ray bright filaments, to the entire extent $\\gtrsim 100\\,{\\rm pc}$ of the nebula. These data also allow us to fully characterize a pair of nearly symmetric, sharp-edged, elongated structures aligned with the orbital axis of the binary system, which lack radio counterparts, but are prominent in very high energy gamma-ray emission. The resulting multifaceted picture of the interaction between energetic outflows and the surrounding medium paves the way for future focused multiwavelength observations and dedicated numerical simulations.",
      "url": "http://arxiv.org/abs/2510.14938v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.14938v1",
      "citations": null,
      "categories": [
        "astro-ph.HE",
        "astro-ph.GA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 423
    },
    {
      "title": "TED++: Submanifold-Aware Backdoor Detection via Layerwise Tubular-Neighbourhood Screening",
      "authors": [
        "Nam Le",
        "Leo Yu Zhang",
        "Kewen Liao",
        "Shirui Pan",
        "Wei Luo"
      ],
      "abstract": "As deep neural networks power increasingly critical applications, stealthy backdoor attacks, where poisoned training inputs trigger malicious model behaviour while appearing benign, pose a severe security risk. Many existing defences are vulnerable when attackers exploit subtle distance-based anomalies or when clean examples are scarce. To meet this challenge, we introduce TED++, a submanifold-aware framework that effectively detects subtle backdoors that evade existing defences. TED++ begins by constructing a tubular neighbourhood around each class's hidden-feature manifold, estimating its local ``thickness'' from a handful of clean activations. It then applies Locally Adaptive Ranking (LAR) to detect any activation that drifts outside the admissible tube. By aggregating these LAR-adjusted ranks across all layers, TED++ captures how faithfully an input remains on the evolving class submanifolds. Based on such characteristic ``tube-constrained'' behaviour, TED++ flags inputs whose LAR-based ranking sequences deviate significantly. Extensive experiments are conducted on benchmark datasets and tasks, demonstrating that TED++ achieves state-of-the-art detection performance under both adaptive-attack and limited-data scenarios. Remarkably, even with only five held-out examples per class, TED++ still delivers near-perfect detection, achieving gains of up to 14\\% in AUROC over the next-best method. The code is publicly available at https://github.com/namle-w/TEDpp.",
      "url": "http://arxiv.org/abs/2510.14299v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.14299v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 424
    },
    {
      "title": "When Flatness Does (Not) Guarantee Adversarial Robustness",
      "authors": [
        "Nils Philipp Walter",
        "Linara Adilova",
        "Jilles Vreeken",
        "Michael Kamp"
      ],
      "abstract": "Despite their empirical success, neural networks remain vulnerable to small, adversarial perturbations. A longstanding hypothesis suggests that flat minima, regions of low curvature in the loss landscape, offer increased robustness. While intuitive, this connection has remained largely informal and incomplete. By rigorously formalizing the relationship, we show this intuition is only partially correct: flatness implies local but not global adversarial robustness. To arrive at this result, we first derive a closed-form expression for relative flatness in the penultimate layer, and then show we can use this to constrain the variation of the loss in input space. This allows us to formally analyze the adversarial robustness of the entire network. We then show that to maintain robustness beyond a local neighborhood, the loss needs to curve sharply away from the data manifold. We validate our theoretical predictions empirically across architectures and datasets, uncovering the geometric structure that governs adversarial vulnerability, and linking flatness to model confidence: adversarial examples often lie in large, flat regions where the model is confidently wrong. Our results challenge simplified views of flatness and provide a nuanced understanding of its role in robustness.",
      "url": "http://arxiv.org/abs/2510.14231v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.14231v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 425
    },
    {
      "title": "Manifold Decoders: A Framework for Generative Modeling from Nonlinear Embeddings",
      "authors": [
        "Riddhish Thakare",
        "Kingdom Mutala Akugri"
      ],
      "abstract": "Classical nonlinear dimensionality reduction (NLDR) techniques like t-SNE, Isomap, and LLE excel at creating low-dimensional embeddings for data visualization but fundamentally lack the ability to map these embeddings back to the original high-dimensional space. This one-way transformation limits their use in generative applications. This paper addresses this critical gap by introducing a system- atic framework for constructing neural decoder architectures for prominent NLDR methods, enabling bidirectional mapping for the first time. We extend this framework by implementing a diffusion-based generative process that operates directly within these learned manifold spaces. Through experiments on the CelebA dataset, we evaluate the reconstruction and generative performance of our approach against autoencoder and standard diffusion model baselines. Our findings reveal a fundamental trade- off: while the decoders successfully reconstruct data, their quality is surpassed by end-to-end optimized autoencoders. Moreover, manifold-constrained diffusion yields poor-quality samples, suggesting that the discrete and sparse nature of classical NLDR embeddings is ill-suited for the continuous inter- polation required by generative models. This work highlights the inherent challenges in retrofitting generative capabilities onto NLDR methods designed primarily for visualization and analysis.",
      "url": "http://arxiv.org/abs/2510.13622v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.13622v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 426
    },
    {
      "title": "Toward Hyper-Dimensional Connectivity in Beyond 6G: A Conceptual Framework",
      "authors": [
        "Ekram Hossain",
        "Angelo Vera-Rivera"
      ],
      "abstract": "Cellular wireless networks enable mobile broadband connectivity for Internet-based applications through their radio access and core network infrastructure. While Fifth-Generation (5G) cellular systems are currently being deployed, ongoing research on cellular technologies primarily focuses on Sixth-Generation (6G) networks to set the stage for developing standards for these systems. Therefore, the time has come to articulate the visions for beyond 6G (B6G) systems. In this article, we present a visionary framework toward hyper-dimensional connectivity in B6G that enables wireless access to hyper-immersive Internet technologies. Our contributions include a conceptual framework for B6G cellular systems with jointly integrated communication, cognition, computing, and cyber-physical capabilities as core connectivity dimensions, a set of technical definitions outlining potential use cases and system-level requirements, a mapping of prospective technology enablers, and a forward-looking research agenda for B6G systems. The conceptual discussions in this article would be helpful for identifying innovation drivers, shaping long-term technical goals, and defining research agendas for the future of mobile broadband technologies.",
      "url": "http://arxiv.org/abs/2510.12896v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.12896v1",
      "citations": null,
      "categories": [
        "cs.NI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 427
    },
    {
      "title": "Cautious Weight Decay",
      "authors": [
        "Lizhang Chen",
        "Jonathan Li",
        "Kaizhao Liang",
        "Baiyu Su",
        "Cong Xie",
        "Nuo Wang Pierse",
        "Chen Liang",
        "Ni Lao",
        "Qiang Liu"
      ],
      "abstract": "We introduce Cautious Weight Decay (CWD), a one-line, optimizer-agnostic modification that applies weight decay only to parameter coordinates whose signs align with the optimizer update. Unlike standard decoupled decay, which implicitly optimizes a regularized or constrained objective, CWD preserves the original loss and admits a bilevel interpretation: it induces sliding-mode behavior upon reaching the stationary manifold, allowing it to search for locally Pareto-optimal stationary points of the unmodified objective. In practice, CWD is a drop-in change for optimizers such as AdamW, Lion, and Muon, requiring no new hyperparameters or additional tuning. For language model pre-training and ImageNet classification, CWD consistently improves final loss and accuracy at million- to billion-parameter scales.",
      "url": "http://arxiv.org/abs/2510.12402v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.12402v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "math.OC",
        "stat.ML"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 428
    },
    {
      "title": "Non-Hermitian Realization of Quantum Dynamics on Embedded Manifolds",
      "authors": [
        "Samuel Alperin"
      ],
      "abstract": "We show that the Floquet Hamiltonian of a quantum particle driven by a general time-periodic imaginary potential is exactly equivalent, at stroboscopic times, to the Hamiltonian of a free particle constrained to a curved Riemannian manifold with fixed embedding. We illustrate the construction for a sinusoidal drive and for the torus of revolution, and outline how the framework can guide experimental design of curved-space quantum dynamics. Our results unify non-Hermitian Floquet physics with spectral geometry and provide a general recipe for engineering quantum dynamics on embedded manifolds.",
      "url": "http://arxiv.org/abs/2510.11845v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.11845v1",
      "citations": null,
      "categories": [
        "quant-ph",
        "math-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 429
    },
    {
      "title": "Cross-correlation of Luminous Red Galaxies with ML-selected AGN in HSC-SSP III: HOD Parameters for Type I and Type II Quasars",
      "authors": [
        "Rodrigo Córdova Rosado",
        "Andy D. Goulding",
        "Jenny E. Greene",
        "Nickolas Kokron",
        "Andrina Nicola",
        "Michael A. Strauss",
        "Ryan C. Hickox"
      ],
      "abstract": "Understanding the dark matter (DM) halo environment in which galaxies that host active galactic nuclei (AGN) reside is a window into the nature of supermassive black hole (SMBH) accretion. We apply halo occupation distribution (HOD) modeling tools to interpret the angular cross-correlation functions between $1.5\\times10^6$ luminous red galaxies (LRGs) and our $\\sim28,500$ Hyper Suprime-Cam + Wide-field Infrared Survey Explorer-selected (and $L_{6 μm}$-limited) AGN to infer the halo properties of distinct quasar samples at physical scales $s>0.1\\,{\\rm Mpc}$, for $z\\in0.7-1.0$. We find that Type I (unobscured) and Type II (obscured) AGN cluster differently, both on small and large physical scales. The derived HODs imply that Type I AGN reside, on average, in substantially ($\\sim3\\times$) more massive halos ($M_h \\sim 10^{13.4} M_\\odot$) than Type II AGN ($M_h \\sim 10^{12.9} M_\\odot$) at $>5σ$ significance. While Type II AGN show one-halo correlations similar to that of galaxies of their average halo mass, the Type I AGN intra-halo clustering signal is significantly shallower. We interpret this observation with HOD methods and find Type I AGN are significantly less likely ($f_{sat}\\sim0.05^{+1}_{-0.05}\\%$) to be found in satellite galaxies than Type II AGN. We find reddened + obscured AGN to have typical satellite fractions for their inferred average halo mass ($\\sim10^{13} M_\\odot$), with $f_{sat} \\sim 20^{+10}_{-5}\\%$. Taken together, these results pose a significant challenge to the strict unified AGN morphological model, and instead suggest that a quasar's spectral class is strongly correlated with its host galaxy's dark matter halo environment. These intriguing results have provided a more complex picture of the SMBH -- DM halo connection, and motivate future analyses of the intrinsic galaxy and accretion properties of AGN.",
      "url": "http://arxiv.org/abs/2510.11780v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.11780v1",
      "citations": null,
      "categories": [
        "astro-ph.GA",
        "astro-ph.CO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 430
    },
    {
      "title": "Renormalization of Interacting Random Graph Models",
      "authors": [
        "Alessio Catanzaro",
        "Diego Garlaschelli",
        "Subodh P. Patil"
      ],
      "abstract": "Random graphs offer a useful mathematical representation of a variety of real world complex networks. Exponential random graphs, for example, are particularly suited towards generating random graphs constrained to have specified statistical moments. In this investigation, we elaborate on a generalization of the former where link probabilities are conditioned on the appearance of other links, corresponding to the introduction of interactions in an effective generalized statistical mechanical formalism. When restricted to the simplest non-trivial case of pairwise interactions, one can derive a closed form renormalization group transformation for maximum coordination number two on the corresponding line graph. Higher coordination numbers do not admit exact closed form renormalization group transformations, a feature that paraphrases the usual absence of exact transformations in two or more dimensional lattice systems. We introduce disorder and study the induced renormalization group flow on its probability assignments, highlighting its formal equivalence to time reversed anisotropic drift-diffusion on the statistical manifold associated with the effective Hamiltonian. We discuss the implications of our findings, stressing the long wavelength irrelevance of certain classes of pair-wise conditioning on random graphs, and conclude with possible applications. These include modeling the scaling behavior of preferential effects on social networks, opinion dynamics, and reinforcement effects on neural networks, as well as how our findings offer a systematic framework to deal with data limitations in inference and reconstruction problems.",
      "url": "http://arxiv.org/abs/2510.07186v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.07186v2",
      "citations": null,
      "categories": [
        "cond-mat.stat-mech",
        "cond-mat.dis-nn",
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 431
    },
    {
      "title": "Stable Robot Motions on Manifolds: Learning Lyapunov-Constrained Neural Manifold ODEs",
      "authors": [
        "David Boetius",
        "Abdelrahman Abdelnaby",
        "Ashok Kumar",
        "Stefan Leue",
        "Abdalla Swikir",
        "Fares J. Abu-Dakka"
      ],
      "abstract": "Learning stable dynamical systems from data is crucial for safe and reliable robot motion planning and control. However, extending stability guarantees to trajectories defined on Riemannian manifolds poses significant challenges due to the manifold's geometric constraints. To address this, we propose a general framework for learning stable dynamical systems on Riemannian manifolds using neural ordinary differential equations. Our method guarantees stability by projecting the neural vector field evolving on the manifold so that it strictly satisfies the Lyapunov stability criterion, ensuring stability at every system state. By leveraging a flexible neural parameterisation for both the base vector field and the Lyapunov function, our framework can accurately represent complex trajectories while respecting manifold constraints by evolving solutions directly on the manifold. We provide an efficient training strategy for applying our framework and demonstrate its utility by solving Riemannian LASA datasets on the unit quaternion (S^3) and symmetric positive-definite matrix manifolds, as well as robotic motions evolving on \\mathbb{R}^3 \\times S^3. We demonstrate the performance, scalability, and practical applicability of our approach through extensive simulations and by learning robot motions in a real-world experiment.",
      "url": "http://arxiv.org/abs/2510.05707v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.05707v1",
      "citations": null,
      "categories": [
        "cs.RO",
        "cs.LG",
        "math.OC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 432
    },
    {
      "title": "From News to Returns: A Granger-Causal Hypergraph Transformer on the Sphere",
      "authors": [
        "Anoushka Harit",
        "Zhongtian Sun",
        "Jongmin Yu"
      ],
      "abstract": "We propose the Causal Sphere Hypergraph Transformer (CSHT), a novel architecture for interpretable financial time-series forecasting that unifies \\emph{Granger-causal hypergraph structure}, \\emph{Riemannian geometry}, and \\emph{causally masked Transformer attention}. CSHT models the directional influence of financial news and sentiment on asset returns by extracting multivariate Granger-causal dependencies, which are encoded as directional hyperedges on the surface of a hypersphere. Attention is constrained via angular masks that preserve both temporal directionality and geometric consistency. Evaluated on S\\&P 500 data from 2018 to 2023, including the 2020 COVID-19 shock, CSHT consistently outperforms baselines across return prediction, regime classification, and top-asset ranking tasks. By enforcing predictive causal structure and embedding variables in a Riemannian manifold, CSHT delivers both \\emph{robust generalisation across market regimes} and \\emph{transparent attribution pathways} from macroeconomic events to stock-level responses. These results suggest that CSHT is a principled and practical solution for trustworthy financial forecasting under uncertainty.",
      "url": "http://arxiv.org/abs/2510.04357v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.04357v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "q-fin.CP"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 433
    },
    {
      "title": "Integrated Planning and Control on Manifolds: Factor Graph Representation and Toolkit",
      "authors": [
        "Peiwen Yang",
        "Weisong Wen",
        "Runqiu Yang",
        "Yuanyuan Zhang",
        "Jiahao Hu",
        "Yingming Chen",
        "Naigui Xiao",
        "Jiaqi Zhao"
      ],
      "abstract": "Model predictive control (MPC) faces significant limitations when applied to systems evolving on nonlinear manifolds, such as robotic attitude dynamics and constrained motion planning, where traditional Euclidean formulations struggle with singularities, over-parameterization, and poor convergence. To overcome these challenges, this paper introduces FactorMPC, a factor-graph based MPC toolkit that unifies system dynamics, constraints, and objectives into a modular, user-friendly, and efficient optimization structure. Our approach natively supports manifold-valued states with Gaussian uncertainties modeled in tangent spaces. By exploiting the sparsity and probabilistic structure of factor graphs, the toolkit achieves real-time performance even for high-dimensional systems with complex constraints. The velocity-extended on-manifold control barrier function (CBF)-based obstacle avoidance factors are designed for safety-critical applications. By bridging graphical models with safety-critical MPC, our work offers a scalable and geometrically consistent framework for integrated planning and control. The simulations and experimental results on the quadrotor demonstrate superior trajectory tracking and obstacle avoidance performance compared to baseline methods. To foster research reproducibility, we have provided open-source implementation offering plug-and-play factors.",
      "url": "http://arxiv.org/abs/2510.04278v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.04278v1",
      "citations": null,
      "categories": [
        "cs.RO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 434
    },
    {
      "title": "Asymmetric rational reductions of 2D-Toda hierarchy and a generalized Frobenius manifold",
      "authors": [
        "Haonan Qu",
        "Qiulan Zhao"
      ],
      "abstract": "We study the local bihamiltonian structures of the asymmetric rational reductions of the 2D-Toda hierarchy (RR2T) of types $(2,1)$ and $(1,2)$ at the full-dispersive level, and construct a three-dimensional generalized Frobenius manifold with non-flat unity associated with the $(2,1)$-type. Furthermore, we explicitly relate the $(2,1)$-type RR2T to the bi-graded Toda and constrained KP hierarchies via linear reciprocal and Miura-type transformations.",
      "url": "http://arxiv.org/abs/2510.04151v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.04151v1",
      "citations": null,
      "categories": [
        "nlin.SI",
        "math-ph",
        "math.DG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 435
    },
    {
      "title": "Efficient Manifold-Constrained Neural ODE for High-Dimensional Datasets",
      "authors": [
        "Muhao Guo",
        "Haoran Li",
        "Yang Weng"
      ],
      "abstract": "Neural ordinary differential equations (NODE) have garnered significant attention for their design of continuous-depth neural networks and the ability to learn data/feature dynamics. However, for high-dimensional systems, estimating dynamics requires extensive calculations and suffers from high truncation errors for the ODE solvers. To address the issue, one intuitive approach is to consider the non-trivial topological space of the data distribution, i.e., a low-dimensional manifold. Existing methods often rely on knowledge of the manifold for projection or implicit transformation, restricting the ODE solutions on the manifold. Nevertheless, such knowledge is usually unknown in realistic scenarios. Therefore, we propose a novel approach to explore the underlying manifold to restrict the ODE process. Specifically, we employ a structure-preserved encoder to process data and find the underlying graph to approximate the manifold. Moreover, we propose novel methods to combine the NODE learning with the manifold, resulting in significant gains in computational speed and accuracy. Our experimental evaluations encompass multiple datasets, where we compare the accuracy, number of function evaluations (NFEs), and convergence speed of our model against existing baselines. Our results demonstrate superior performance, underscoring the effectiveness of our approach in addressing the challenges of high-dimensional datasets.",
      "url": "http://arxiv.org/abs/2510.04138v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.04138v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 436
    },
    {
      "title": "The Principle of Isomorphism: A Theory of Population Activity in Grid Cells and Beyond",
      "authors": [
        "Maoshen Xu",
        "Fei Song",
        "Yuxiu Shao",
        "Bailu Si",
        "Shanshan Qin"
      ],
      "abstract": "Identifying the principles that determine neural population activity is paramount in the field of neuroscience. We propose the Principle of Isomorphism (PIso): population activity preserves the essential mathematical structures of the tasks it supports. Using grid cells as a model system, we show that the neural metric task is characterized by a flat Riemannian manifold, while path integration is characterized by an Abelian Lie group. We prove that each task independently constrains population activity to a toroidal topology. We further show that these perspectives are unified naturally in Euclidean space, where commutativity and flatness are intrinsically compatible and can be extended to related systems including head-direction cells and 3D grid cells. To examine how toroidal topology maps onto single-cell firing patterns, we develop a minimal network architecture that explicitly constrains population activity to toroidal manifolds. Our model robustly generates hexagonal firing fields and reveals systematic relationships between network parameters and grid spacings. Crucially, we demonstrate that conformal isometry, a commonly proposed hypothesis, alone is insufficient for hexagonal field formation. Our findings establish a direct link between computational tasks and the hexagonal-toroidal organization of grid cells, thereby providing a general framework for understanding population activity in neural systems and designing task-informed architectures in machine learning.",
      "url": "http://arxiv.org/abs/2510.02853v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.02853v2",
      "citations": null,
      "categories": [
        "q-bio.NC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 437
    },
    {
      "title": "Action Deviation-Aware Inference for Low-Latency Wireless Robots",
      "authors": [
        "Jeyoung Park",
        "Yeonsub Lim",
        "Seungeun Oh",
        "Jihong Park",
        "Jinho Choi",
        "Seong-Lyun Kim"
      ],
      "abstract": "To support latency-sensitive AI applications ranging from autonomous driving to industrial robot manipulation, 6G envisions distributed ML with computational resources in mobile, edge, and cloud connected over hyper-reliable low-latency communication (HRLLC). In this setting, speculative decoding can facilitate collaborative inference of models distributively deployed: a lightweight on-device model locally generates drafts while a more capable remote target model on a server verifies and corrects them in parallel with speculative sampling, thus resulting in lower latency without compromising accuracy. However, unlike autoregressive text generation, behavior cloning policies, typically used for embodied AI applications, cannot parallelize verification and correction for multiple drafts as each generated action depends on observation updated by a previous action. To this end, we propose Action Deviation-Aware Hybrid Inference (ADAHI), wherein drafts are selectively transmitted and verified based on action deviation, which has a strong correlation with action's rejection probability by the target model. By invoking server operation only when necessary, communication and computational overhead can be reduced while accuracy gain from speculative sampling is preserved. Experiments on our testbed show that ADAHI reduces transmission and server operations by approximately 40%, lowers end-to-end latency by 39.2%, and attains up to 97.2% of the task-success rate of baseline that invokes speculative sampling for every draft embedding vector.",
      "url": "http://arxiv.org/abs/2510.02851v2",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.02851v2",
      "citations": null,
      "categories": [
        "cs.RO",
        "cs.DC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 438
    },
    {
      "title": "Topological Invariance and Breakdown in Learning",
      "authors": [
        "Yongyi Yang",
        "Tomaso Poggio",
        "Isaac Chuang",
        "Liu Ziyin"
      ],
      "abstract": "We prove that for a broad class of permutation-equivariant learning rules (including SGD, Adam, and others), the training process induces a bi-Lipschitz mapping between neurons and strongly constrains the topology of the neuron distribution during training. This result reveals a qualitative difference between small and large learning rates $η$. With a learning rate below a topological critical point $η^*$, the training is constrained to preserve all topological structure of the neurons. In contrast, above $η^*$, the learning process allows for topological simplification, making the neuron manifold progressively coarser and thereby reducing the model's expressivity. Viewed in combination with the recent discovery of the edge of stability phenomenon, the learning dynamics of neuron networks under gradient descent can be divided into two phases: first they undergo smooth optimization under topological constraints, and then enter a second phase where they learn through drastic topological simplifications. A key feature of our theory is that it is independent of specific architectures or loss functions, enabling the universal application of topological methods to the study of deep learning.",
      "url": "http://arxiv.org/abs/2510.02670v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.02670v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 439
    },
    {
      "title": "StelLA: Subspace Learning in Low-rank Adaptation using Stiefel Manifold",
      "authors": [
        "Zhizhong Li",
        "Sina Sajadmanesh",
        "Jingtao Li",
        "Lingjuan Lyu"
      ],
      "abstract": "Low-rank adaptation (LoRA) has been widely adopted as a parameter-efficient technique for fine-tuning large-scale pre-trained models. However, it still lags behind full fine-tuning in performance, partly due to its insufficient exploitation of the geometric structure underlying low-rank manifolds. In this paper, we propose a geometry-aware extension of LoRA that uses a three-factor decomposition $U\\!SV^\\top$. Analogous to the structure of singular value decomposition (SVD), it separates the adapter's input and output subspaces, $V$ and $U$, from the scaling factor $S$. Our method constrains $U$ and $V$ to lie on the Stiefel manifold, ensuring their orthonormality throughout the training. To optimize on the Stiefel manifold, we employ a flexible and modular geometric optimization design that converts any Euclidean optimizer to a Riemannian one. It enables efficient subspace learning while remaining compatible with existing fine-tuning pipelines. Empirical results across a wide range of downstream tasks, including commonsense reasoning, math and code generation, image classification, and image generation, demonstrate the superior performance of our approach against the recent state-of-the-art variants of LoRA. Code is available at https://github.com/SonyResearch/stella.",
      "url": "http://arxiv.org/abs/2510.01938v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2510.01938v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 440
    },
    {
      "title": "Sparse view tomographic reconstruction of elongated objects using learned primal-dual networks",
      "authors": [
        "Buda Bajić",
        "Johannes A. J. Huber",
        "Benedikt Neyses",
        "Linus Olofsson",
        "Ozan Öktem"
      ],
      "abstract": "",
      "url": "https://openalex.org/W4392538367",
      "year": 2025,
      "venue": "Engineering Applications of Artificial Intelligence",
      "source": "openalex",
      "doi": "10.1016/j.engappai.2025.112295",
      "pdf_url": "https://doi.org/10.1016/j.engappai.2025.112295",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 441
    },
    {
      "title": "S2PW-Mamba: Pinwheel and Wavelet-based Spatial-Spectral Mamba for Hyperspectral Image Classification",
      "authors": [
        "Lianhui Liang",
        "Wangli He",
        "Ying Zhang",
        "Y. J. Zeng",
        "Thomas Wu",
        "Antonio Plaza"
      ],
      "abstract": "",
      "url": "https://openalex.org/W4413966806",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.36227/techrxiv.175693604.45516380/v1",
      "pdf_url": "https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.175693604.45516380/v1",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 442
    },
    {
      "title": "Introduction",
      "authors": [
        "Yidong Xu",
        "Jianghong Mao",
        "Weijie Zhuge",
        "Xiaoniu Yu",
        "Ping Wu"
      ],
      "abstract": "",
      "url": "https://openalex.org/W4413860807",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.1007/978-981-96-8237-9_1",
      "pdf_url": "https://link.springer.com/content/pdf/10.1007/978-981-96-8237-9_1.pdf",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 443
    },
    {
      "title": "SparseFraudNet: A Graph-based Approach for Cold-start Fraud Detection with Information Aggregation",
      "authors": [
        "Wen Zhang",
        "Rui Li",
        "Quan Bai",
        "Song Wang"
      ],
      "abstract": "Online reviews play a critical role in influencing consumer’s purchasing decision on e-commerce, making them a prime target for manipulation through fraudulent reviews. Although various Fraud Detection (FD) techniques have been presented, a crucial problem still remains unaddressed, i.e., the cold-start problem in FD, which refers to the difficulty in identifying fraudulent reviews due to limited historical data for new users and new products. Existing graph-based detection methods, while effective for well-connected nodes, are suffering Sparse Graph (SG) connections in cold-start FD. In this paper, we propose a novel approach called SparseFraudNet to address the problem of cold-start FD with information aggregation. Specifically, the local information aggregation is proposed to dynamically optimize neighbor selection using Reinforcement Learning (RL) with Bernoulli Multi-Armed Bandit (BMAB), with the goal to capture the five key types of relations among reviews. The global information aggregation is proposed to leverage Graph Coarsening (GC) with manifold learning and spectral clustering to mitigate adjacency matrix sparsity for new users under new products using Sparse Spectral Clustering (SSC). Experiments on the YelpZip-Cold and YelpNYC-Cold datasets demonstrate that the proposed SparseFraudNet approach significantly outperforms state-of-the-art methods in FD in terms of accuracy, precision, recall, F1 measure and AUC to identify fraudulent reviews of new users under new ",
      "url": "https://openalex.org/W4413418666",
      "year": 2025,
      "venue": "ACM Transactions on Information Systems",
      "source": "openalex",
      "doi": "10.1145/3748719",
      "pdf_url": "https://dl.acm.org/doi/pdf/10.1145/3748719",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 444
    },
    {
      "title": "Hyperbolic Deep Learning for Foundation Models: A Survey",
      "authors": [
        "Neil He",
        "Hiren Madhu",
        "Ngoc Bui",
        "Meng‐Lin Yang",
        "Rex Ying"
      ],
      "abstract": "Foundation models pre-trained on massive datasets, including large language models (LLMs), vision-language models (VLMs), and large multimodal models, have demonstrated remarkable success in diverse downstream tasks. However, recent studies have shown fundamental limitations of these models: (1) limited representational capacity, (2) lower adaptability, and (3) diminishing scalability. These shortcomings raise a critical question: is Euclidean geometry truly the optimal inductive bias for all foundation models, or could incorporating alternative geometric spaces enable models to better align with the intrinsic structure of real-world data and improve reasoning processes? Hyperbolic spaces, a class of non-Euclidean manifolds characterized by exponential volume growth with respect to distance, offer a mathematically grounded solution. These spaces enable low-distortion embeddings of hierarchical structures (e.g., trees, taxonomies) and power-law distributions with substantially fewer dimensions compared to Euclidean counterparts. Recent advances have leveraged these properties to enhance foundation models, including improving LLMs' complex reasoning ability, VLMs' zero-shot generalization, and cross-modal semantic alignment, while maintaining parameter efficiency. This paper provides a comprehensive review of hyperbolic neural networks and their recent development for foundation models. We further outline key challenges and research directions to advance the field.",
      "url": "https://openalex.org/W4412875482",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.1145/3711896.3736564",
      "pdf_url": "https://dl.acm.org/doi/pdf/10.1145/3711896.3736564",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 445
    },
    {
      "title": "OptWake-YOLO: a lightweight and efficient ship wake detection model based on optical remote sensing images",
      "authors": [
        "Robert C. Qiu",
        "Nan Bi"
      ],
      "abstract": "Introduction Ship wakes exhibit more distinctive characteristics than vessels themselves, making wake detection more feasible than direct ship detection. However, challenges persist due to sea surface interference, meteorological conditions, and coastal structures, while practical applications demand lightweight models with fast detection speeds. Methods We propose OptWake-YOLO, a lightweight ship wake detection model with three key innovations: A RepConv-based RCEA module in the Backbone combining efficient layer aggregation with reparameterization to enhance feature extraction. An Adaptive Dynamic Feature Fusion Network (ADFFN) in the Neck integrating channel attention with Dynamic Upsampling (Dysample). A Shared Lightweight Object Detection Head (SLODH) using parameter sharing and Group Normalization. Results Experiments on the SWIM dataset show OptWake-YOLO improves mAP50 by 1.5% (to 93.2%) and mAP50-95 by 2.9% (to 66.5%) compared to YOLOv11n, while reducing parameters by 40.7% (to 1.6M) and computation by 25.8% (to 4.9 GFLOPs), maintaining 303 FPS speed. Discussion The model demonstrates superior performance in complex maritime conditions through: RCEA's multi-branch feature extraction. ADFFN's adaptive multi-scale fusion. SLODH's efficient detection architecture. Ablation studies confirm each component's contribution to balancing accuracy and efficiency for real-time wake detection.",
      "url": "https://openalex.org/W4412821756",
      "year": 2025,
      "venue": "Frontiers in Marine Science",
      "source": "openalex",
      "doi": "10.3389/fmars.2025.1624323",
      "pdf_url": "https://www.frontiersin.org/journals/marine-science/articles/10.3389/fmars.2025.1624323/pdf",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 446
    },
    {
      "title": "A Method for Multimodal Remote Sensing Image Classification",
      "authors": [
        "Zhong Sun",
        "Bin Hu"
      ],
      "abstract": "In remote sensing, images are widely used in applications, such as land cover classification, urban monitoring, and disaster management, providing rich information about the Earth's surface. However, due to data heterogeneity and scarcity, different modalities of remote-sensing images often face challenges in classification tasks. The proposed deep learning model for remote-sensing image classification addresses these challenges through multimodal fusion. By combining a convolutional neural network, a generative adversarial network, and a graph convolutional network, the model is organized into three main components: data preprocessing and feature extraction, multimodal data generation and enhancement, and multimodal feature fusion and classification. Experimental results on the Hyperspectral-Light Detection and Ranging Houston2013 dataset and the Hyperspectral-Synthetic Aperture Radar Berlin dataset show that the proposed method significantly outperforms traditional methods and other deep learning models in classification performance, with better stability and robustness.",
      "url": "https://openalex.org/W4412083045",
      "year": 2025,
      "venue": "Journal of Organizational and End User Computing",
      "source": "openalex",
      "doi": "10.4018/joeuc.384397",
      "pdf_url": "https://www.igi-global.com/ViewTitle.aspx?TitleId=384397&isxn=9798337311579",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 447
    },
    {
      "title": "Ensemble Kalman methods: A mean-field perspective",
      "authors": [
        "Edoardo Calvello",
        "Stephanie Reich",
        "Andrew M. Stuart"
      ],
      "abstract": "Ensemble Kalman methods, introduced in 1994 in the context of ocean state estimation, are now widely used for state estimation and parameter estimation (inverse problems) in many arenae. Their success stems from the fact that they take an underlying computational model as a black box to provide a systematic, derivative-free methodology for incorporating observations; furthermore the ensemble approach allows for sensitivities and uncertainties to be calculated. Analysis of the accuracy of ensemble Kalman methods, especially in terms of uncertainty quantification, is lagging behind empirical success; this paper provides a unifying mean-field-based framework for their analysis. Both state estimation and parameter estimation problems are considered, and formulations in both discrete and continuous time are employed. For state estimation problems, both the control and filtering approaches are considered; analogously for parameter estimation problems, the optimization and Bayesian perspectives are both studied. As well as providing an elegant framework, the mean-field perspective also allows for the derivation of a variety of methods used in practice. In addition it unifies a wide-ranging literature in the field and suggests open problems.",
      "url": "https://openalex.org/W4411917706",
      "year": 2025,
      "venue": "Acta Numerica",
      "source": "openalex",
      "doi": "10.1017/s0962492924000060",
      "pdf_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/94C9B874BBD4F11B8D36DD42D9F01BC7/S0962492924000060a.pdf/div-class-title-ensemble-kalman-methods-a-mean-field-perspective-div.pdf",
      "citations": 6,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 448
    },
    {
      "title": "Talks",
      "authors": [],
      "abstract": "",
      "url": "https://openalex.org/W4412764722",
      "year": 2025,
      "venue": "FEBS Open Bio",
      "source": "openalex",
      "doi": "10.1002/2211-5463.70069",
      "pdf_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/2211-5463.70069",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 449
    },
    {
      "title": "Destructive Creation of New Invasive Technologies: Generative Artificial Intelligence Behaviour",
      "authors": [
        "Mario Coccia"
      ],
      "abstract": "This study proposes a new concept that explains a source of technological change: The invasive behaviour of general purpose technologies that breaks into scientific and technological ecosystems with accelerated diffusion of new products and processes that destroy the usage value of all units previously used. This study highlights the dynamics of the invasive destruction of new path-breaking technologies in driving innovative activity. Invasive technologies conquer the scientific, technological, and business spaces of alternative technologies by introducing manifold radical innovations that support technological, economic, and social change. The proposed theoretical framework is verified empirically in new technologies of neural network architectures, comparing transformer technology (a deep learning architecture having unsupervised and semi-supervised algorithms that create new contents and mimic human ability, supporting Generative Artificial Intelligence) to Long Short-Term Memory (LSTM) and Recurrent Neural Networks (RNNs). Statistical evidence here, based on patent analyses, reveals that the exponential growth rate of transformer technology over a period of five years (2020–2024) is 45.91% more than double compared to the alternative technologies of LSTM (21.17%) and RNN (18.15%). Moreover, the proposed invasive rate in technological space shows that is very high for transformer technology at the level of 2.2%, whereas for LSTM it is 1.39% and for RNN it is 1.22% over 202",
      "url": "https://openalex.org/W4411495197",
      "year": 2025,
      "venue": "Technologies",
      "source": "openalex",
      "doi": "10.3390/technologies13070261",
      "pdf_url": "https://www.mdpi.com/2227-7080/13/7/261/pdf?version=1750412465",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 450
    },
    {
      "title": "Nonlinear Dynamics in Game Theory as a New Mathematical Approach to Analysing Strategic Behaviour",
      "authors": [
        "Abiodun Finbarrs Oketunji"
      ],
      "abstract": "This research presents a novel mathematical framework integrating nonlinear dynamics with game theory to analyse strategic behaviour in complex multi-agent systems. Traditional game-theoretic approaches often assume equilibrium convergence and rational decision-making, yet empirical observations reveal persistent oscillations, chaotic behaviour, and multi-stability in strategic interactions. We develop a unified theory incorporating bifurcation analysis, strange attractors, and Lyapunov stability to characterise the full spectrum of dynamical behaviours in strategic settings. Our framework introduces the concept of strategic bifurcations—qualitative changes in equilibrium structure induced by parameter variations in payoff functions or behavioural rules. We establish conditions for Hopf bifurcations in replicator dynamics, derive analytical expressions for limit cycle amplitudes, and characterise routes to chaos through period-doubling cascades. The theory extends to n-player games with heterogeneous learning rates, revealing that chaos becomes increasingly prevalent as system complexity grows. We prove that the basin of attraction for stable Nash equilibria shrinks exponentially withthe number of players, whilst the measure of chaotic regimes expands. Applications to evolutionary biology, financial markets, and social dynamics demonstratethe framework’s predictive power. Our results challenge the primacy of equilibrium analysis in game theory and establish nonlinear dynamics",
      "url": "https://openalex.org/W4411528337",
      "year": 2025,
      "venue": "Preprints.org",
      "source": "openalex",
      "doi": "10.20944/preprints202506.1702.v1",
      "pdf_url": "https://www.preprints.org/frontend/manuscript/5447ba31dca22ca010cb741da18287e1/download_pub",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 451
    },
    {
      "title": "sHGCN: Simplified hyperbolic graph convolutional neural networks",
      "authors": [
        "Paredes Arévalo",
        "Alexis Molina",
        "Álvaro Ciudad"
      ],
      "abstract": "Hyperbolic geometry has emerged as a powerful tool for modeling complex, structured data, particularly where hierarchical or tree-like relationships are present. By enabling embeddings with lower distortion, hyperbolic neural networks offer promising alternatives to Euclidean-based models for capturing intricate data structures. Despite these advantages, they often face performance challenges, particularly in computational efficiency and tasks requiring high precision. In this work, we address these limitations by simplifying key operations within hyperbolic neural networks, achieving notable improvements in both runtime and performance. Our findings demonstrate that streamlined hyperbolic operations can lead to substantial gains in computational speed and predictive accuracy, making hyperbolic neural networks a more viable choice for a broader range of applications.",
      "url": "https://openalex.org/W4415311858",
      "year": 2025,
      "venue": "arXiv (Cornell University)",
      "source": "openalex",
      "doi": "10.48550/arxiv.2506.14438",
      "pdf_url": "https://arxiv.org/pdf/2506.14438",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 452
    },
    {
      "title": "The Past, Present and Future of the Corporate Actor: Ontological, Epistemological and Theoretical Considerations",
      "authors": [
        "Michaela Haase",
        "Elke Schuessler",
        "Ute Schmiel",
        "Günther Ortmann",
        "Andreas Suchanek",
        "Dennis Schoeneborn"
      ],
      "abstract": "Abstract Corporate actors are more resourceful, more powerful and more capable of influencing their own conditions of action than most other actors. This curated article argues that to imagine the future of corporate actors in a world that is rapidly changing due to the possibilities of digital technology and sustainability challenges, we also need to revisit dominant conceptualizations of the corporate actor in economic or social science research. The four essays included in this article draw on different theories to elaborate on the role that corporate actors play, could play, or should play in the future. The problems that the authors identify in their essays include the power and wrongdoing of corporate actors; the ability of corporate actors to create their environment, not just react to it; the lack of a concept of (corporate) responsibility capable of responding to the desiderata of society; and the accountability of corporate actorhood outside the boundaries of formal organizations. By providing reflections on these problems, this article—written against the background of the historical peculiarities of German business administration research marked by a neglect of collectivist concepts such as corporate actors—shows that research and society actively co-construct the roles and responsibilities of corporate actors. The performativity of theories addressing corporate actors is thus a concern for business administration scholars aiming to contribute to a more sustainabl",
      "url": "https://openalex.org/W4411342480",
      "year": 2025,
      "venue": "Schmalenbach Journal of Business Research",
      "source": "openalex",
      "doi": "10.1007/s41471-025-00213-w",
      "pdf_url": "https://link.springer.com/content/pdf/10.1007/s41471-025-00213-w.pdf",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 453
    },
    {
      "title": "Fault-Tolerant Logical Measurements via Homological Measurement",
      "authors": [
        "Benjamin Ide",
        "Manoj G. Gowda",
        "Priya J. Nadkarni",
        "Guillaume Dauphinais"
      ],
      "abstract": "We introduce homological measurement, a framework for measuring the logical Pauli operators encoded in Calderbank-Shor-Steane stabilizer codes. The framework is based on the algebraic description of such codes as chain complexes. Protocols such as lattice surgery and some of its recent generalizations are shown to be special cases of homological measurement. Using this framework, we develop a specific protocol called edge expanded homological measurement for fault-tolerant measurement of arbitrary logical Pauli operators of general quantum low density parity-check codes, requiring a number of ancillary qubits growing only linearly with the weight of the logical operator measured, and guarantee that the distance of the code is preserved. We further benchmark our protocol numerically in a photonic architecture based on Gottesman-Kitaev-Preskill qubits, showing that the logical error rates of various codes are on par with other methods requiring more ancilla qubits.",
      "url": "https://openalex.org/W4403884303",
      "year": 2025,
      "venue": "Physical Review X",
      "source": "openalex",
      "doi": "10.1103/physrevx.15.021088",
      "pdf_url": "http://link.aps.org/pdf/10.1103/PhysRevX.15.021088",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 454
    },
    {
      "title": "Artificial Intelligence Across Borders: Transforming Industries Through Intelligent Innovation",
      "authors": [],
      "abstract": "",
      "url": "https://openalex.org/W4411150579",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.70593/978-93-49910-25-6",
      "pdf_url": "https://deepscienceresearch.com/dsr/catalog/download/156/832/1717",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 455
    },
    {
      "title": "Algorithm- and Data-Dependent Generalization Bounds for Score-Based Generative Models",
      "authors": [
        "Benjamin Dupuis",
        "Dario Shariatian",
        "Maxime Haddouche",
        "Alain Durmus",
        "Umut Şimşekli"
      ],
      "abstract": "Score-based generative models (SGMs) have emerged as one of the most popular classes of generative models. A substantial body of work now exists on the analysis of SGMs, focusing either on discretization aspects or on their statistical performance. In the latter case, bounds have been derived, under various metrics, between the true data distribution and the distribution induced by the SGM, often demonstrating polynomial convergence rates with respect to the number of training samples. However, these approaches adopt a largely approximation theory viewpoint, which tends to be overly pessimistic and relatively coarse. In particular, they fail to fully explain the empirical success of SGMs or capture the role of the optimization algorithm used in practice to train the score network. To support this observation, we first present simple experiments illustrating the concrete impact of optimization hyperparameters on the generalization ability of the generated distribution. Then, this paper aims to bridge this theoretical gap by providing the first algorithmic- and data-dependent generalization analysis for SGMs. In particular, we establish bounds that explicitly account for the optimization dynamics of the learning algorithm, offering new insights into the generalization behavior of SGMs. Our theoretical findings are supported by empirical results on several datasets.",
      "url": "https://openalex.org/W4416073882",
      "year": 2025,
      "venue": "arXiv (Cornell University)",
      "source": "openalex",
      "doi": "10.48550/arxiv.2506.03849",
      "pdf_url": "https://arxiv.org/pdf/2506.03849",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 456
    },
    {
      "title": "Vector Ising spin annealer for minimizing Ising Hamiltonians",
      "authors": [
        "James Cummins",
        "Natalia G. Berloff"
      ],
      "abstract": "Abstract Complex optimization problems can be solved via dedicated machines which encode the problem in the couplings of spin Hamiltonians. However, traditional physical minimizers often select excited states due to limitations in spin dynamics. We introduce the Vector Ising Spin Annealer (VISA), a framework in gain-based computing that leverages light-matter interactions. We show that VISA overcomes the limitations by enabling spins to operate within a three-dimensional space, thereby providing a robust solution for effectively minimizing Ising Hamiltonians. Our comparative analysis demonstrates VISA’s superior performance relative to conventional single-dimension spin optimizers, highlighting its capacity to surmount significant energy barriers in intricate landscapes. Detailed studies on cyclic and random graphs reveal VISA’s proficiency in dynamically evolving the energy landscape through time-dependent gain and penalty annealing, underscoring its potential in advancing the field of complex problem-solving in physics-inspired and physics-based computing.",
      "url": "https://openalex.org/W4410855184",
      "year": 2025,
      "venue": "Communications Physics",
      "source": "openalex",
      "doi": "10.1038/s42005-025-02145-7",
      "pdf_url": "https://www.nature.com/articles/s42005-025-02145-7.pdf",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 457
    },
    {
      "title": "Optimal Rotational Smoothing on S1: Why Only the Poisson Kernel Survives on the Circle",
      "authors": [
        "D. R. Stanley"
      ],
      "abstract": "Circular signals—angles, phases, orientations—pervade modern science and engineering yet resist straightforward linear smoothing techniques.Context. Circular data arise in disciplines ranging from wind forecasting to phase-unwrapping and cryo-EM.Problem. Despite a century of practice, no consensus exists on which smoothing kernel on the unit circle achieves the optimal balance between symmetry, stability and spectral localisation.Method. We formulate six axioms—(1) reality &amp;amp; evenness, (2) unit mass, (3) an analytic strip with simple poles, (4) a single inflection, (5) positive-definiteness, and (6) a half-height bandwidth criterion—and fuse contour integration, Paley–Wiener theory and Bochner–Herglotz positivity to analyse their joint implications. A residue calculation converts the analytic-strip constraint into an exponential Fourier envelope, while positivity confines residue phases, and normalisation plus curvature conditions fix the remaining scalar.Result. The only kernel satisfying all six axioms is the Poisson (order-1 Butterworth) family Kₐ(φ) = sinh a ⁄ (cosh a − cos φ), a &amp;gt; 0,uniquely determined up to its bandwidth parameter.Implications. Numerical experiments in denoising, spectral-leakage suppression and wind-direction forecasting confirm that once the axioms are accepted, practitioners should “pick their a, not their window,” and adopt Kₐ as the default circular smoother.",
      "url": "https://openalex.org/W4410807039",
      "year": 2025,
      "venue": "",
      "source": "openalex",
      "doi": "10.31219/osf.io/a4gvz_v1",
      "pdf_url": "https://osf.io/a4gvz_v1/download",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 458
    },
    {
      "title": "SpaceTimePilot: Generative Rendering of Dynamic Scenes Across Space and Time",
      "authors": [
        "Zhening Huang",
        "Hyeonho Jeong",
        "Xuelin Chen",
        "Yulia Gryaditskaya",
        "Tuanfeng Y. Wang",
        "Joan Lasenby",
        "Chun-Hao Huang"
      ],
      "abstract": "We present SpaceTimePilot, a video diffusion model that disentangles space and time for controllable generative rendering. Given a monocular video, SpaceTimePilot can independently alter the camera viewpoint and the motion sequence within the generative process, re-rendering the scene for continuous and arbitrary exploration across space and time. To achieve this, we introduce an effective animation time-embedding mechanism in the diffusion process, allowing explicit control of the output video's motion sequence with respect to that of the source video. As no datasets provide paired videos of the same dynamic scene with continuous temporal variations, we propose a simple yet effective temporal-warping training scheme that repurposes existing multi-view datasets to mimic temporal differences. This strategy effectively supervises the model to learn temporal control and achieve robust space-time disentanglement. To further enhance the precision of dual control, we introduce two additional components: an improved camera-conditioning mechanism that allows altering the camera from the first frame, and CamxTime, the first synthetic space-and-time full-coverage rendering dataset that provides fully free space-time video trajectories within a scene. Joint training on the temporal-warping scheme and the CamxTime dataset yields more precise temporal control. We evaluate SpaceTimePilot on both real-world and synthetic data, demonstrating clear space-time disentanglement and strong results compared to prior work. Project page: https://zheninghuang.github.io/Space-Time-Pilot/ Code: https://github.com/ZheningHuang/spacetimepilot",
      "url": "http://arxiv.org/abs/2512.25075v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25075v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.AI",
        "cs.RO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 459
    },
    {
      "title": "Scaling Open-Ended Reasoning to Predict the Future",
      "authors": [
        "Nikhil Chandak",
        "Shashwat Goel",
        "Ameya Prabhu",
        "Moritz Hardt",
        "Jonas Geiping"
      ],
      "abstract": "High-stakes decision making involves reasoning under uncertainty about the future. In this work, we train language models to make predictions on open-ended forecasting questions. To scale up training data, we synthesize novel forecasting questions from global events reported in daily news, using a fully automated, careful curation recipe. We train the Qwen3 thinking models on our dataset, OpenForesight. To prevent leakage of future information during training and evaluation, we use an offline news corpus, both for data generation and retrieval in our forecasting system. Guided by a small validation set, we show the benefits of retrieval, and an improved reward function for reinforcement learning (RL). Once we obtain our final forecasting system, we perform held-out testing between May to August 2025. Our specialized model, OpenForecaster 8B, matches much larger proprietary models, with our training improving the accuracy, calibration, and consistency of predictions. We find calibration improvements from forecasting training generalize across popular benchmarks. We open-source all our models, code, and data to make research on language model forecasting broadly accessible.",
      "url": "http://arxiv.org/abs/2512.25070v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25070v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.CL"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 460
    },
    {
      "title": "Many Minds from One Model: Bayesian Transformers for Population Intelligence",
      "authors": [
        "Diji Yang",
        "Yi Zhang"
      ],
      "abstract": "Despite their scale and success, modern transformers are almost universally trained as single-minded systems: optimization produces one deterministic set of parameters, representing a single functional hypothesis about the data. Motivated by the idea that intelligence emerge from many minds, we propose Population Bayesian Transformers (B-Trans), which transform a standard Large Language Model into a Bayesian Transformer model to supports sampling diverse yet coherent model instances from a single set of pre-trained weights.\n  B-Trans introduces a Bayesian-motivated posterior proxy by treating the bias-like offsets in normalization layers as stochastic variables with a Gaussian variational approximation, inducing a distribution over model behavior without the cost of training full Bayesian neural networks. Sampling from this proxy yields a set of model instances with diverse behaviors while maintaining general competence. To preserve coherence within each generation, we freeze the sampled noise at the sequence level, enforcing temporal consistency across tokens. B-Trans allows for population-level decision-making, where aggregating predictions across sampled individuals significantly enhances exploration. Experiments across zero-shot generation, Reinforcement Learning with Verifiable Rewards (RLVR), and RL without explicit labels demonstrate that B-Trans effectively leverage the wisdom of crowds, yielding superior semantic diversity while achieving better task performance compared to deterministic baselines.",
      "url": "http://arxiv.org/abs/2512.25063v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25063v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.CL"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 461
    },
    {
      "title": "Melting curve of correlated iron at Earth's core conditions from machine-learned DFT+DMFT",
      "authors": [
        "Rishi Rao",
        "Li Zhu"
      ],
      "abstract": "Reliable constraints on iron's melting curve at Earth's inner-core boundary require accurate finite-temperature electronic correlations, yet DFT+DMFT calculations remain too costly for large-scale thermodynamic sampling. Here, we develop a machine-learning accelerator for charge self-consistent DFT+DMFT by training E(3)-equivariant graph neural networks to predict the local self-energy and Fermi level from atomic environments, providing an efficient warm start to the DMFT self-consistency loop. Using high-throughput data for Fe, FeO, and NiO, we obtain a 2-4 times reuduction in DMFT iterations. Leveraging this improvement, we generate correlated energies and forces for Fe at core pressures, train a neural-network interatomic potential, and determine the melting curve via two-phase coexistence simulations. We obtain a predicted melting temperature of 6225 K at 330 GPa.",
      "url": "http://arxiv.org/abs/2512.25061v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25061v1",
      "citations": null,
      "categories": [
        "cond-mat.mtrl-sci",
        "physics.geo-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 462
    },
    {
      "title": "On the geometry and topology of representations: the manifolds of modular addition",
      "authors": [
        "Gabriela Moisescu-Pareja",
        "Gavin McCracken",
        "Harley Wiltzer",
        "Vincent Létourneau",
        "Colin Daniels",
        "Doina Precup",
        "Jonathan Love"
      ],
      "abstract": "The Clock and Pizza interpretations, associated with architectures differing in either uniform or learnable attention, were introduced to argue that different architectural designs can yield distinct circuits for modular addition. In this work, we show that this is not the case, and that both uniform attention and trainable attention architectures implement the same algorithm via topologically and geometrically equivalent representations. Our methodology goes beyond the interpretation of individual neurons and weights. Instead, we identify all of the neurons corresponding to each learned representation and then study the collective group of neurons as one entity. This method reveals that each learned representation is a manifold that we can study utilizing tools from topology. Based on this insight, we can statistically analyze the learned representations across hundreds of circuits to demonstrate the similarity between learned modular addition circuits that arise naturally from common deep learning paradigms.",
      "url": "http://arxiv.org/abs/2512.25060v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25060v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 463
    },
    {
      "title": "Reliable and Resilient Collective Communication Library for LLM Training and Serving",
      "authors": [
        "Wei Wang",
        "Nengneng Yu",
        "Sixian Xiong",
        "Zaoxing Liu"
      ],
      "abstract": "Modern ML training and inference now span tens to tens of thousands of GPUs, where network faults can waste 10--15\\% of GPU hours due to slow recovery. Common network errors and link fluctuations trigger timeouts that often terminate entire jobs, forcing expensive checkpoint rollback during training and request reprocessing during inference. We present R$^2$CCL, a fault-tolerant communication library that provides lossless, low-overhead failover by exploiting multi-NIC hardware. R$^2$CCL performs rapid connection migration, bandwidth-aware load redistribution, and resilient collective algorithms to maintain progress under failures. We evaluate R$^2$CCL on two 8-GPU H100 InfiniBand servers and via large-scale ML simulators modeling hundreds of GPUs with diverse failure patterns. Experiments show that R$^2$CCL is highly robust to NIC failures, incurring less than 1\\% training and less than 3\\% inference overheads. R$^2$CCL outperforms baselines AdapCC and DejaVu by 12.18$\\times$ and 47$\\times$, respectively.",
      "url": "http://arxiv.org/abs/2512.25059v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25059v1",
      "citations": null,
      "categories": [
        "cs.DC",
        "cs.LG",
        "cs.NI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 464
    },
    {
      "title": "Fluid dynamics as intersection problem",
      "authors": [
        "Nikita Nekrasov",
        "Paul Wiegmann"
      ],
      "abstract": "We formulate the covariant hydrodynamics equations describing the fluid dynamics as the problem of intersection theory on the infinite dimensional symplectic manifold associated with spacetime. This point of view separates the structures related to the equation of state, the geometry of spacetime, and structures related to the (differential) topology of spacetime. We point out a five-dimensional origin of the formalism of Lichnerowicz and Carter. Our formalism also incorporates the chiral anomaly and Onsager quantization. We clarify the relation between the canonical velocity and Landau $4$-velocity, the meaning of Kelvin's theorem. Finally, we discuss some connections to topological strings, Poisson sigma models, and topological field theories in various dimensions.",
      "url": "http://arxiv.org/abs/2512.25053v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25053v1",
      "citations": null,
      "categories": [
        "hep-th",
        "gr-qc",
        "math-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 465
    },
    {
      "title": "Bilinear tau forms of quantum Painlevé equations and $\\mathbb{C}^2/\\mathbb{Z}_2$ blowup relations in SUSY gauge theories",
      "authors": [
        "Giulio Bonelli",
        "Anton Shchechkin",
        "Alessandro Tanzini"
      ],
      "abstract": "We derive bilinear tau forms of the canonically quantized Painlevé equations, thereby relating them to those previously obtained from the $\\mathbb{C}^2/\\mathbb{Z}_2$ blowup relations for the $\\mathcal{N}=2$ supersymmetric gauge theory partition functions on a general $Ω$-background. We fully fix the refined Painlevé/gauge theory dictionary by formulating the proper equations for the quantum nonautonomous Painlevé Hamiltonians. We also describe the symmetry structure of the quantum Painlevé tau functions and, as a byproduct of this analysis, obtain the $\\mathbb{C}^2/\\mathbb{Z}_2$ blowup relations in the nontrivial holonomy sector of the gauge theory.",
      "url": "http://arxiv.org/abs/2512.25051v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25051v1",
      "citations": null,
      "categories": [
        "math-ph",
        "hep-th",
        "math.QA",
        "nlin.SI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 466
    },
    {
      "title": "Arithmetic with spatiotemporal optical vortex of integer and fractional topological charges",
      "authors": [
        "Hsiao-Chih Huang",
        "Chen-Ting Liao",
        "Hui Min Leung"
      ],
      "abstract": "Spatiotemporal optical vortices carry transverse orbital angular momentum (t-OAM), which give rise to spatiotemporal topological charge (ST-TC). To unleash the full potential of t-OAM in expanding the capacity of communication and computing, we demonstrate the first optical information-processing pipeline capable of performing addition and subtraction on ST-TC values, regardless of whether they are integer or fractional. Additionally, we established a readout method for those mathematical operations through imaging spectral analysis, providing a robust optical basis toward arithmetic operations and verification. These new capabilities mark crucial advancements toward full arithmetic operations on the ST-TC of light for bosonic state computation and information processing.",
      "url": "http://arxiv.org/abs/2512.25049v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25049v1",
      "citations": null,
      "categories": [
        "physics.optics",
        "physics.data-an"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 467
    },
    {
      "title": "Generative Classifiers Avoid Shortcut Solutions",
      "authors": [
        "Alexander C. Li",
        "Ananya Kumar",
        "Deepak Pathak"
      ],
      "abstract": "Discriminative approaches to classification often learn shortcuts that hold in-distribution but fail even under minor distribution shift. This failure mode stems from an overreliance on features that are spuriously correlated with the label. We show that generative classifiers, which use class-conditional generative models, can avoid this issue by modeling all features, both core and spurious, instead of mainly spurious ones. These generative classifiers are simple to train, avoiding the need for specialized augmentations, strong regularization, extra hyperparameters, or knowledge of the specific spurious correlations to avoid. We find that diffusion-based and autoregressive generative classifiers achieve state-of-the-art performance on five standard image and text distribution shift benchmarks and reduce the impact of spurious correlations in realistic applications, such as medical or satellite datasets. Finally, we carefully analyze a Gaussian toy setting to understand the inductive biases of generative classifiers, as well as the data properties that determine when generative classifiers outperform discriminative ones.",
      "url": "http://arxiv.org/abs/2512.25034v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25034v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "cs.AI",
        "cs.CV",
        "cs.NE"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 468
    },
    {
      "title": "Testing Monotonicity in a Finite Population",
      "authors": [
        "Jiafeng Chen",
        "Jonathan Roth",
        "Jann Spiess"
      ],
      "abstract": "We consider the extent to which we can learn from a completely randomized experiment whether everyone has treatment effects that are weakly of the same sign, a condition we call monotonicity. From a classical sampling perspective, it is well-known that monotonicity is untestable. By contrast, we show from the design-based perspective -- in which the units in the population are fixed and only treatment assignment is stochastic -- that the distribution of treatment effects in the finite population (and hence whether monotonicity holds) is formally identified. We argue, however, that the usual definition of identification is unnatural in the design-based setting because it imagines knowing the distribution of outcomes over different treatment assignments for the same units. We thus evaluate the informativeness of the data by the extent to which it enables frequentist testing and Bayesian updating. We show that frequentist tests can have nontrivial power against some alternatives, but power is generically limited. Likewise, we show that there exist (non-degenerate) Bayesian priors that never update about whether monotonicity holds. We conclude that, despite the formal identification result, the ability to learn about monotonicity from data in practice is severely limited.",
      "url": "http://arxiv.org/abs/2512.25032v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25032v1",
      "citations": null,
      "categories": [
        "econ.EM",
        "math.ST",
        "stat.ME"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 469
    },
    {
      "title": "On Nonlinear Inertial Transformations",
      "authors": [
        "Nicholas Agia"
      ],
      "abstract": "It is often assumed that the most general transformation between two inertial reference frames is affine linear in their Cartesian coordinates, an assumption which is however not true. We provide a complete derivation of the most general inertial frame transformation, which is indeed nonlinear; along the way, we shall find that the conditions of preserving the Law of Inertia take the form of Schwarzian differential equations, providing perhaps the simplest possible physics setting in which the Schwarzian derivative appears. We then demonstrate that the most general such inertial transformation which further preserves the speed of light in all directions is, however, still affine linear. Physically, this paper may be viewed as a reduction of the number of postulates needed to uniquely specify special relativity by one, as well as a proof that inertial transformations automatically imbue spacetime with a vector space structure, albeit in one higher dimension than might be expected. Mathematically, this paper may be viewed as a derivation of the higher-dimensional analog of the Schwarzian differential equation and its most general solution.",
      "url": "http://arxiv.org/abs/2512.25024v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25024v1",
      "citations": null,
      "categories": [
        "physics.class-ph",
        "gr-qc"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 470
    },
    {
      "title": "ResponseRank: Data-Efficient Reward Modeling through Preference Strength Learning",
      "authors": [
        "Timo Kaufmann",
        "Yannick Metz",
        "Daniel Keim",
        "Eyke Hüllermeier"
      ],
      "abstract": "Binary choices, as often used for reinforcement learning from human feedback (RLHF), convey only the direction of a preference. A person may choose apples over oranges and bananas over grapes, but which preference is stronger? Strength is crucial for decision-making under uncertainty and generalization of preference models, but hard to measure reliably. Metadata such as response times and inter-annotator agreement can serve as proxies for strength, but are often noisy and confounded. We propose ResponseRank to address the challenge of learning from noisy strength signals. Our method uses relative differences in proxy signals to rank responses to pairwise comparisons by their inferred preference strength. To control for systemic variation, we compare signals only locally within carefully constructed strata. This enables robust learning of utility differences consistent with strength-derived rankings while making minimal assumptions about the strength signal. Our contributions are threefold: (1) ResponseRank, a novel method that robustly learns preference strength by leveraging locally valid relative strength signals; (2) empirical evidence of improved sample efficiency and robustness across diverse tasks: synthetic preference learning (with simulated response times), language modeling (with annotator agreement), and RL control tasks (with simulated episode returns); and (3) the Pearson Distance Correlation (PDC), a novel metric that isolates cardinal utility learning from ordinal accuracy.",
      "url": "http://arxiv.org/abs/2512.25023v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25023v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 471
    },
    {
      "title": "Detector Response Matrices, Effective Areas, and Flash-Effective Areas for Radiation Detectors",
      "authors": [
        "Gregory Bowers",
        "Eve Chase",
        "William Ford",
        "Daniel Coupland",
        "Brian Larsen",
        "Caleb Roecker",
        "Karl Smith",
        "Kurtis Bartlett",
        "Katherine Gattiker Katherine Mesick"
      ],
      "abstract": "A Detector Response Matrix (DRM) is a discrete representation of an instrument's Detector Response Function (DRF), which quantifies how many discrete energy depositions occur in a detector volume for a given distribution of particles incident on the detector. For simple radiation detectors that can count such energy depositions (such as scintillators, Proportional Counter Tubes (PCTs), etc), we consider the ideal counting DRF, $\\mathbf{G}_\\varphi (E_\\mathrm{in}, E_\\mathrm{dep})$, which relates the detector's counting histogram (number of energy depositions within a given channel) to an incident particles characterization, $\\varphi$ (e.g. incident flux, fluence, intensity). From the counting DRF we can derive the counting DRM, the effective area, and the flash effective area (which measures the total energy deposited in the detector from a large, instantaneous fluence).",
      "url": "http://arxiv.org/abs/2512.25021v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25021v1",
      "citations": null,
      "categories": [
        "physics.ins-det",
        "hep-ex",
        "math-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 472
    },
    {
      "title": "Convergence of the generalization error for deep gradient flow methods for PDEs",
      "authors": [
        "Chenguang Liu",
        "Antonis Papapantoleon",
        "Jasper Rou"
      ],
      "abstract": "The aim of this article is to provide a firm mathematical foundation for the application of deep gradient flow methods (DGFMs) for the solution of (high-dimensional) partial differential equations (PDEs). We decompose the generalization error of DGFMs into an approximation and a training error. We first show that the solution of PDEs that satisfy reasonable and verifiable assumptions can be approximated by neural networks, thus the approximation error tends to zero as the number of neurons tends to infinity. Then, we derive the gradient flow that the training process follows in the ``wide network limit'' and analyze the limit of this flow as the training time tends to infinity. These results combined show that the generalization error of DGFMs tends to zero as the number of neurons and the training time tend to infinity.",
      "url": "http://arxiv.org/abs/2512.25017v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25017v1",
      "citations": null,
      "categories": [
        "math.NA",
        "cs.LG",
        "q-fin.CP",
        "stat.ML"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 473
    },
    {
      "title": "A note on semistable unitary operators on $L^2(\\mathbb{R})$",
      "authors": [
        "Xianghong Chen"
      ],
      "abstract": "In this note, we present a characterization of semistable unitary operators on $L^2(\\mathbb{R})$, under the assumption that the operator is (i) translation-invariant, (ii) symmetric, and (iii) locally uniformly continuous (LUC) under dilation. As a consequence, we characterize one-parameter groups formed by such operators, which are of the form $e^{iβt|{d}/{dx}|^α}$, with $α,β\\in\\mathbb R$.",
      "url": "http://arxiv.org/abs/2512.25013v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25013v1",
      "citations": null,
      "categories": [
        "math.FA",
        "math-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 474
    },
    {
      "title": "FoundationSLAM: Unleashing the Power of Depth Foundation Models for End-to-End Dense Visual SLAM",
      "authors": [
        "Yuchen Wu",
        "Jiahe Li",
        "Fabio Tosi",
        "Matteo Poggi",
        "Jin Zheng",
        "Xiao Bai"
      ],
      "abstract": "We present FoundationSLAM, a learning-based monocular dense SLAM system that addresses the absence of geometric consistency in previous flow-based approaches for accurate and robust tracking and mapping. Our core idea is to bridge flow estimation with geometric reasoning by leveraging the guidance from foundation depth models. To this end, we first develop a Hybrid Flow Network that produces geometry-aware correspondences, enabling consistent depth and pose inference across diverse keyframes. To enforce global consistency, we propose a Bi-Consistent Bundle Adjustment Layer that jointly optimizes keyframe pose and depth under multi-view constraints. Furthermore, we introduce a Reliability-Aware Refinement mechanism that dynamically adapts the flow update process by distinguishing between reliable and uncertain regions, forming a closed feedback loop between matching and optimization. Extensive experiments demonstrate that FoundationSLAM achieves superior trajectory accuracy and dense reconstruction quality across multiple challenging datasets, while running in real-time at 18 FPS, demonstrating strong generalization to various scenarios and practical applicability of our method.",
      "url": "http://arxiv.org/abs/2512.25008v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25008v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 475
    },
    {
      "title": "Grassmannian Geometries for Non-Planar On-Shell Diagrams",
      "authors": [
        "Artyom Lisitsyn",
        "Umut Oktem",
        "Melissa Sherman-Bennett",
        "Jaroslav Trnka"
      ],
      "abstract": "On-shell diagrams are gauge invariant quantities which play an important role in the description of scattering amplitudes. Based on the principles of generalized unitarity, they are given by products of elementary three-point amplitudes where the kinematics of internal on-shell legs are determined by cut conditions. In the ${\\cal N}=4$ Super Yang-Mills (SYM) theory, the dual formulation for on-shell diagrams produces the same quantities as canonical forms on the Grassmannian $G(k,n)$. Most of the work in this direction has been devoted to the planar diagrams, which dominate in the large $N$ limit of gauge theories. On the mathematical side, planar on-shell diagrams correspond to cells of the positive Grassmannian $G_+(k,n)$ which have been very extensively studied in the literature in the past 20 years. In this paper, we focus on the non-planar on-shell diagrams which are relevant at finite $N$. In particular, we use the triplet formulation of Maximal-Helicity-Violating (MHV) on-shell diagrams to obtain certain regions in the Grassmannian $G(2,n)$. These regions are unions of positive Grassmannians with different orderings (referred to as oriented regions). We explore the features of these unions, and show that they are pseudo-positive geometries, in contrast to positive geometry of a single oriented region. For all non-planar diagrams which are \\emph{internally planar} there always exists a strongly connected geometry, and for those that are \\emph{irreducible}, there exists a geometry with no spurious facets. We also prove that the already known identity moves, square and sphere moves, form the complete set of identity moves for all MHV on-shell diagrams.",
      "url": "http://arxiv.org/abs/2512.25005v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25005v1",
      "citations": null,
      "categories": [
        "hep-th",
        "math.CO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 476
    },
    {
      "title": "Efficiently Estimating Data Efficiency for Language Model Fine-tuning",
      "authors": [
        "Gyung Hyun Je",
        "Colin Raffel"
      ],
      "abstract": "While large language models (LLMs) demonstrate reasonable zero-shot capability across many downstream tasks, fine-tuning is a common practice to improve their performance. However, a task's data efficiency--i.e., the number of fine-tuning examples needed to achieve a desired level of performance--is often unknown, resulting in costly cycles of incremental annotation and retraining. Indeed, we demonstrate across a curated set of 30 specialized tasks that performant LLMs may struggle zero-shot but can attain stronger performance after fine-tuning. This motivates the need for methods to predict a task's data efficiency without requiring incremental annotation. After introducing a concrete metric that quantifies a task's data efficiency, we propose using the gradient cosine similarity of low-confidence examples to predict data efficiency based on a small number of labeled samples. We validate our approach on a diverse set of tasks with varying data efficiencies, attaining 8.6% error in overall data efficiency prediction and typically eliminating hundreds of unnecessary annotations on each task. Our experiment results and implementation code are available on GitHub.",
      "url": "http://arxiv.org/abs/2512.24991v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24991v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 477
    },
    {
      "title": "Wall crossing, string networks and quantum toroidal algebras",
      "authors": [
        "Yegor Zenkevich"
      ],
      "abstract": "We investigate BPS states in 4d N=4 supersymmetric Yang-Mills theory and the corresponding (p, q) string networks in Type IIB string theory. We propose a new interpretation of the algebra of line operators in this theory as a tensor product of vector representations of a quantum toroidal algebra, which determines protected spin characters of all framed BPS states. We identify the SL(2,Z)-noninvariant choice of the coproduct in the quantum toroidal algebra with the choice of supersymmetry subalgebra preserved by the BPS states and interpret wall crossing operators as Drinfeld twists of the coproduct. Kontsevich-Soibelman spectrum generator is then identified with Khoroshkin-Tolstoy universal R-matrix.",
      "url": "http://arxiv.org/abs/2512.24988v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24988v1",
      "citations": null,
      "categories": [
        "hep-th",
        "math-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 478
    },
    {
      "title": "Local approximations of global Hamiltonian from inclusion of algebras",
      "authors": [
        "Yidong Chen",
        "Nima Lashkari",
        "Kwing Lam Leung"
      ],
      "abstract": "We write down the global Hamiltonian of conformal field theory (CFT) in finite volume in terms of the modular Hamiltonian of the vacuum reduced to a local ball-shaped region, and use it to propose local approximations to the global Minkowski Hamiltonian in quantum field theory (QFT). The proposed Hamiltonians are motivated by the operator-algebraic property of nuclearity. They are constructed from the characteristic functions of inclusion of algebras and can be viewed as regulators of the modular Hamiltonian of local algebras of QFT",
      "url": "http://arxiv.org/abs/2512.25062v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25062v1",
      "citations": null,
      "categories": [
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 479
    },
    {
      "title": "The variety of orthogonal frames",
      "authors": [
        "Laura Casabella",
        "Alessio Sammartano"
      ],
      "abstract": "An orthogonal n-frame is an ordered set of n pairwise orthogonal vectors. The set of all orthogonal n-frames in a d-dimensional quadratic vector space is an algebraic variety V(d,n). In this paper, we investigate the variety V(d,n) as well as the quadratic ideal I(d,n) generated by the orthogonality relations, which cuts out V(d,n). We classify the irreducible components of V(d,n), give criteria for the ideal I(d,n) to be prime or a complete intersection, and for the variety V(d,n) to be normal. We also give near-equivalent conditions for V(d,n) to be factorial. Applications are given to the theory of Lovász-Saks-Schrijver ideals.",
      "url": "http://arxiv.org/abs/2512.25058v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25058v1",
      "citations": null,
      "categories": [
        "math.AC",
        "math.AG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 480
    },
    {
      "title": "Emergence of 3D Superconformal Ising Criticality on the Fuzzy Sphere",
      "authors": [
        "Yin Tang",
        "Cristian Voinea",
        "Liangdong Hu",
        "Zlatko Papić",
        "W. Zhu"
      ],
      "abstract": "Supersymmetric conformal field theories (SCFTs) form a unique subset of quantum field theories which provide powerful insights into strongly coupled critical phenomena. Here, we present a microscopic and non-perturbative realization of the three-dimensional $\\mathcal{N}=1$ superconformal Ising critical point, based on a Yukawa-type coupling between a 3D Ising CFT and a gauged Majorana fermion. Using the recently developed fuzzy sphere regularization, we directly extract the scaling dimensions of low-lying operators via the state-operator correspondence. At the critical point, we demonstrate conformal multiplet structure together with the hallmark of emergent spacetime supersymmetry through characteristic relations between fermionic and bosonic operators. Moreover, by tuning the Yukawa coupling, we explicitly track the evolution of operator spectra from the decoupled Ising-Majorana fixed point to the interacting superconformal fixed point, revealing renormalization-group flow at the operator level. Our results establish a controlled, non-perturbative microscopic route to 3D SCFTs.",
      "url": "http://arxiv.org/abs/2512.25054v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25054v1",
      "citations": null,
      "categories": [
        "cond-mat.str-el",
        "cond-mat.stat-mech",
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 481
    },
    {
      "title": "Amplitude constraints on dark energy",
      "authors": [
        "Scott Melville"
      ],
      "abstract": "This talk gives a short introduction to the ``UV/EFT correspondence\", which uses scattering amplitudes to relate the Effective Field Theory (EFT) coefficients probed by low-energy measurements to properties of the underlying high-energy (UV) completion. This includes recent ``positivity bounds\" on EFT coefficients, which are the low-energy signatures of causality and unitarity in the UV. To illustrate their phenomenological impact, I apply these bounds to a simple EFT for dark energy and compare with recent cosmological observations.",
      "url": "http://arxiv.org/abs/2512.25047v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": "10.58027/3q8k-ew90",
      "pdf_url": "https://arxiv.org/pdf/2512.25047v1",
      "citations": null,
      "categories": [
        "gr-qc"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 482
    },
    {
      "title": "The Hochschild homology of a noncommutative symmetric quotient stack",
      "authors": [
        "Rina Anno",
        "Vladimir Baranovsky",
        "Timothy Logvinenko"
      ],
      "abstract": "We prove an orbifold type decomposition theorem for the Hochschild homology of the symmetric powers of a small DG category $\\mathcal{A}$. In noncommutative geometry, these can be viewed as the noncommutative symmetric quotient stacks of $\\mathcal{A}$. We use this decomposition to show that the total Hochschild homology of the symmetric powers of $\\mathcal{A}$ is isomorphic to the symmetric algebra $S^*(\\mathrm{HH}_\\bullet(\\mathcal{A}) \\otimes t \\mathbb{k}[t])$. Our methods are explicit - we construct mutually inverse homotopy equivalences of the standard Hochschild complexes involved. These explicit maps are then used to induce from the symmetric algebra onto the total Hochschild homology the structures of the Fock space for the Heisenberg algebra of $\\mathcal{A}$, of a Hopf algebra, and of a free $λ$-ring generated by $\\mathrm{HH}_\\bullet(\\mathcal{A})$.",
      "url": "http://arxiv.org/abs/2512.25039v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25039v1",
      "citations": null,
      "categories": [
        "math.AG",
        "math.CT",
        "math.RT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 483
    },
    {
      "title": "Anomalous (3+1)d Fermionic Topological Quantum Field Theories via Symmetry Extension",
      "authors": [
        "Zheyan Wan",
        "Juven Wang"
      ],
      "abstract": "Discrete finite-group global symmetries may suffer from nonperturbative 't-Hooft anomalies. Such global anomalies can be canceled by anomalous symmetry-preserving topological quantum field theories (TQFTs), which contain no local point operators but only extended excitations such as line and surface operators. In this work, we study mixed gauge-gravitational nonperturbative global anomalies of Weyl fermions (or Weyl semimetals in condensed matter) charged under discrete Abelian internal symmetries in four-dimensional spacetime, with spacetime-internal fermionic symmetry $G=$Spin$\\times_{\\mathbb{Z}_2^{\\rm F}}\\mathbb{Z}_{2m}^{\\rm F}$ or Spin$\\times\\mathbb{Z}_n$ that contains fermion parity $\\mathbb{Z}_{2}^{\\rm F}$. We determine the minimal finite gauge group $K$ of anomalous $G$-symmetric TQFTs that can match the fermionic anomaly via the symmetry-extension construction $1 \\to K \\to G_{\\rm Tot} \\to G \\to 1$, where the anomaly in $G$ is trivialized upon pullback to $G_{\\rm Tot}$, computed by Atiyah-Patodi-Singer eta invariant. This allows one to replace a $G$-symmetric four-dimensional Weyl fermion by an anomalous $G$-symmetric discrete-$K$-gauge TQFT as an alternative low-energy theory in the same deformation class. As an application, we show that the four-dimensional Standard Model with 15 Weyl fermions per family, in the absence of a sterile right-handed neutrino $ν_R$, exhibits mixed gauge-gravitational global anomalies between baryon and lepton number symmetries $({\\bf B \\pm L})$ and spacetime diffeomorphisms. We identify the corresponding minimal $K$-gauge fermionic TQFT that cancels these anomalies and can be interpreted as a gapped, topologically ordered dark sector replacing missing Weyl fermions via symmetry extension, without invoking conventional Anderson-Higgs symmetry breaking.",
      "url": "http://arxiv.org/abs/2512.25038v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25038v1",
      "citations": null,
      "categories": [
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 484
    },
    {
      "title": "Mod $p$ Poincaré duality for $p$-adic period domains",
      "authors": [
        "Guillaume Pignon-Ywanne"
      ],
      "abstract": "In this article, we introduce a new class of smooth partially proper rigid analytic varieties over a $p$-adic field that satisfy Poincaré duality for étale cohomology with mod $p$-coefficients : the varieties satisfying \"primitive comparison with compact support\". We show that almost proper varieties, as well as p-adic (weakly admissible) period domains in the sense of Rappoport-Zink belong to this class. In particular, we recover Poincaré duality for almost proper varieties as first established by Li-Reinecke-Zavyalov, and we compute the étale cohomology with $\\mathbb{F}_p$-coefficients of p-adic period domains, generalizing a computation of Colmez-Dospinescu-Niziol for Drinfeld's symmetric spaces. The arguments used in this paper rely crucially on Mann's six functors formalism for solid $\\mathcal{O}^{+,a}/π$ coefficients.",
      "url": "http://arxiv.org/abs/2512.25029v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25029v1",
      "citations": null,
      "categories": [
        "math.AG",
        "math.NT",
        "math.RT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 485
    },
    {
      "title": "Real Riemann Surfaces: Smooth and Discrete",
      "authors": [
        "Johanna Düntsch",
        "Felix Günther"
      ],
      "abstract": "This paper develops a discrete theory of real Riemann surfaces based on quadrilateral cellular decompositions (quad-graphs) and a linear discretization of the Cauchy-Riemann equations. We construct a discrete analogue of an antiholomorphic involution and classify the topological types of discrete real Riemann surfaces, recovering the classical results on the number of real ovals and the separation of the surface.\n  Central to our approach is the construction of a symplectic homology basis adapted to the discrete involution. Using this basis, we prove that the discrete period matrix admits the same canonical decomposition $Π= \\frac{1}{2} H + i T$ as in the smooth setting, where $H$ encodes the topological type and $T$ is purely imaginary. This structural result bridges the gap between combinatorial models and the classical theory of real algebraic curves.",
      "url": "http://arxiv.org/abs/2512.25022v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25022v1",
      "citations": null,
      "categories": [
        "math.CV",
        "math.CO",
        "math.DG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 486
    },
    {
      "title": "Bounding regularity of $\\mathrm{VI}^m$-modules",
      "authors": [
        "Wee Liang Gan",
        "Khoa Ta"
      ],
      "abstract": "Fix a finite field $\\mathbb{F}$. Let $\\mathrm{VI}$ be a skeleton of the category of finite dimensional $\\mathbb{F}$-vector spaces and injective $\\mathbb{F}$-linear maps. We study $\\mathrm{VI}^m$-modules over a noetherian commutative ring in the nondescribing characteristic case. We prove that if a finitely generated $\\mathrm{VI}^m$-module is generated in degree $\\leqslant d$ and related in degree $\\leqslant r$, then its regularity is bounded above by a function of $m$, $d$, and $r$. A key ingredient of the proof is a shift theorem for finitely generated $\\mathrm{VI}^m$-modules.",
      "url": "http://arxiv.org/abs/2512.25010v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25010v1",
      "citations": null,
      "categories": [
        "math.RT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 487
    },
    {
      "title": "The splitting field and generators of the elliptic surface $Y^2=X^3 +t^{360} +1$",
      "authors": [
        "Sajad Salami"
      ],
      "abstract": "The splitting field of an elliptic surface $\\mathcal{E}/\\mathbb{Q}(t)$ is the smallest finite extension $\\mathcal{K} \\subset \\mathbb{C}$ such that all $\\mathbb{C}(t)$-rational points are defined over $\\mathcal{K}(t)$. In this paper, we provide a symbolic algorithmic approach to determine the splitting field and a set of $68$ linearly independent generators for the Mordell--Weil lattice of Shioda's elliptic surface $Y^2=X^3 +t^{360} +1$. This surface is noted for having the largest known rank 68 for an elliptic curve over $\\mathbb{C}(t)$.\n  Our methodology utilizes the known decomposition of the Mordell-Weil Lattice of this surface into Lattices of ten rational elliptic surfaces and one $K3$ surface. We explicitly compute the defining polynomials of the splitting field, which reach degrees of 1728 and 5760, and verify the results via height pairing matrices and specialized symbolic software packages.",
      "url": "http://arxiv.org/abs/2512.25009v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25009v1",
      "citations": null,
      "categories": [
        "math.NT",
        "math.AG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 488
    },
    {
      "title": "Fast Poisson brackets and constraint algebras in canonical gravity",
      "authors": [
        "Will Barker"
      ],
      "abstract": "In the study of alternative or extended theories of gravity, Dirac's Hamiltonian constraint algorithm is invaluable for enumerating the propagating modes and gauge symmetries. For gravity, this canonical approach is frequently applied as a means for finding pathologies such as strongly coupled modes; more generally it facilitates the reconstruction of gauge symmetries and the quantization of gauge theories. For gravity, however, the algorithm can become notoriously arduous to implement. We present a simple computer algebra package for efficiently computing Poisson brackets and reconstructing constraint algebras. The tools are stress-tested against pure general relativity and modified gravity, including the order reduction of general relativity at two loops.",
      "url": "http://arxiv.org/abs/2512.25007v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25007v1",
      "citations": null,
      "categories": [
        "physics.comp-ph",
        "gr-qc"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 489
    },
    {
      "title": "Distributions of wide binary stars in theory and in Gaia data: III. Orbital momenta, masses, and manifestations of MOND",
      "authors": [
        "Valeri V. Makarov"
      ],
      "abstract": "Using the censored catalog of 103,169 resolved Gaia DR3 binary stars with accurate astrometric data for each component, a new observable, object-specific parameter is computed for each pair: the projected orbital momentum. This parameter is the product of four functions of physical characteristics: total mass, semimajor axis, eccentricity, and inclination angle. Using the previously estimated marginal probability densities of eccentricity and semimajor axis, and assuming an isotropic orientation of binary systems, the sample distribution of mass was adjusted using a concordance metric of the observed and synthetic distributions of orbital momenta and an ad hoc functional model. The best-fitting mass density model is found to faithfully reproduce the observed dependence of orbital momenta on apparent separation, although the absolute luminosity distributions indicate a tendency of the widest systems to more frequently include solar-type primaries. The anticipated manifestation of MOND is computed in the investigated parameter space \\{separation, momentum\\}. This effect is absent in the given data. The median total mass of the widest Gaia binaries is found to be somewhat higher than that of the tighter pairs, which is interpreted as a dynamical age effect.",
      "url": "http://arxiv.org/abs/2512.25002v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25002v1",
      "citations": null,
      "categories": [
        "astro-ph.SR"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 490
    },
    {
      "title": "Dissipative corrections to the particle momentum spectrum of a decoupling fluid",
      "authors": [
        "Francesco Becattini",
        "Daniele Roselli",
        "Xin-Li Sheng"
      ],
      "abstract": "We present an \\emph{ab initio} calculation within quantum statistical field theory and linear response theory, of the dissipative correction to the momentum spectrum of scalar particles emitted at decoupling (freeze-out) from a relativistic fluid assuming the initial state to be in local thermodynamic equilibrium. We obtain an expansion of the Wigner function of the interacting quantum field in terms of the gradients of the classical thermo-hydrodynamic fields - four-temperature vector and reduced chemical potential - evaluated on the initial local-equilibrium hypersurface, rather than on the decoupling (freeze-out) hypersurface as usual in kinetic theory. The gradient expansion includes an unexpected zeroth order term depending on the differences between thermo-hydrodynamic fields at the decoupling and the initial hypersurface. This term encodes a memory of the initial state which is related to the long-distance persistence of the correlation function between Wigner operator and stress-energy tensor and charged current that is discussed in detail. We address the phenomenological implications of these corrections for the momentum spectra measured in relativistic nuclear collisions.",
      "url": "http://arxiv.org/abs/2512.24994v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24994v1",
      "citations": null,
      "categories": [
        "nucl-th",
        "cond-mat.stat-mech",
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 491
    },
    {
      "title": "DarkEQA: Benchmarking Vision-Language Models for Embodied Question Answering in Low-Light Indoor Environments",
      "authors": [
        "Yohan Park",
        "Hyunwoo Ha",
        "Wonjun Jo",
        "Tae-Hyun Oh"
      ],
      "abstract": "Vision Language Models (VLMs) are increasingly adopted as central reasoning modules for embodied agents. Existing benchmarks evaluate their capabilities under ideal, well-lit conditions, yet robust 24/7 operation demands performance under a wide range of visual degradations, including low-light conditions at night or in dark environments--a core necessity that has been largely overlooked. To address this underexplored challenge, we present DarkEQA, an open-source benchmark for evaluating EQA-relevant perceptual primitives under multi-level low-light conditions. DarkEQA isolates the perception bottleneck by evaluating question answering from egocentric observations under controlled degradations, enabling attributable robustness analysis. A key design feature of DarkEQA is its physical fidelity: visual degradations are modeled in linear RAW space, simulating physics-based illumination drop and sensor noise followed by an ISP-inspired rendering pipeline. We demonstrate the utility of DarkEQA by evaluating a wide range of state-of-the-art VLMs and Low-Light Image Enhancement (LLIE) models. Our analysis systematically reveals VLMs' limitations when operating under these challenging visual conditions. Our code and benchmark dataset will be released upon acceptance.",
      "url": "http://arxiv.org/abs/2512.24985v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24985v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.AI",
        "cs.LG",
        "cs.RO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 492
    },
    {
      "title": "The Supersymmetry of Cuts in Pure Gauge Theory and Gravity",
      "authors": [
        "Jacob L. Bourjaily"
      ],
      "abstract": "At tree-level, scattering amplitudes involving only gluons or gravitons are unaffected by supersymmetry, allowing them to be efficiently encoded by and extracted from those of maximally supersymmetric (N=4,8) theories. This fails beyond tree-level, of course, but much less than would be expected. We show that all the leading singularities of (sub-maximally or) non-supersymmetric theories can be organized into `generalized' superfunctions, in terms of which all helicity components can be effectively encoded. These functions differ from those of maximally supersymmetric theories by an extent determined by loop-order -- broken into a sum over 2^L supersymmetric pieces.",
      "url": "http://arxiv.org/abs/2512.24984v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24984v1",
      "citations": null,
      "categories": [
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 493
    },
    {
      "title": "Optical Spiking Neural Networks via Rogue-Wave Statistics",
      "authors": [
        "Bahadır Utku Kesgin",
        "Gülsüm Yaren Durdu",
        "Uğur Teğin"
      ],
      "abstract": "Optical computing could reduce the energy cost of artificial intelligence by leveraging the parallelism and propagation speed of light. However, implementing nonlinear activation, essential for machine learning, remains challenging in low-power optical systems dominated by linear wave physics. Here, we introduce an optical spiking neural network that uses optical rogue-wave statistics as a programmable firing mechanism. By establishing a homomorphism between free-space diffraction and neuronal integration, we demonstrate that phase-engineered caustics enable robust, passive thresholding: sparse spatial spikes emerge when the local intensity exceeds a significant-intensity rogue-wave criterion. Using a physics-informed digital twin, we optimize granular phase masks to deterministically concentrate energy into targeted detector regions, enabling end-to-end co-design of the optical transformation and a lightweight electronic readout. We experimentally validate the approach on BreastMNIST and Olivetti Faces, achieving accuracies of 82.45\\% and 95.00\\%, respectively, competitive with standard digital baselines. These results demonstrate that extreme-wave phenomena, often treated as deleterious fluctuations, can be harnessed as structural nonlinearity for scalable, energy-efficient neuromorphic photonic inference.",
      "url": "http://arxiv.org/abs/2512.24983v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24983v1",
      "citations": null,
      "categories": [
        "physics.optics"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 494
    },
    {
      "title": "A Modal Logic for Possibilistic Reasoning with Fuzzy Formal Contexts",
      "authors": [
        "Prosenjit Howlader",
        "Churn-Jung Liau"
      ],
      "abstract": "We introduce a two-sort weighted modal logic for possibilistic reasoning with fuzzy formal contexts. The syntax of the logic includes two types of weighted modal operators corresponding to classical necessity ($\\Box$) and sufficiency ($\\boxminus$) modalities and its formulas are interpreted in fuzzy formal contexts based on possibility theory. We present its axiomatization that is \\emph{sound} with respect to the class of all fuzzy context models. In addition, both the necessity and sufficiency fragments of the logic are also individually complete with respect to the class of all fuzzy context models. We highlight the expressive power of the logic with some illustrative examples. As a formal context is the basic construct of formal concept analysis (FCA), we generalize three main notions in FCA, i.e., formal concepts, object oriented concepts, and property oriented concepts, to their corresponding $c$-cut concepts in fuzzy formal contexts. Then, we show that our logical language can represent all three of these generalized notions. Finally, we demonstrate the possibility of extending our logic to reasoning with multi-relational fuzzy contexts, in which the Boolean combinations of different fuzzy relations are allowed.",
      "url": "http://arxiv.org/abs/2512.24980v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24980v1",
      "citations": null,
      "categories": [
        "cs.LO",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 495
    },
    {
      "title": "Semiclassics, branes, and extremality",
      "authors": [
        "Adolfo Holguin"
      ],
      "abstract": "We revisit the problem of computing extremal and non-extremal three point functions of semiclassical probes with single trace operators and point out certain inconsistencies in previous approaches in the literature. We clarify the roles of wavefunctions and averaging over moduli, concluding that holographic computations may be performed with or without averaging. By carefully implementing the extrapolate dictionary for extremal correlators we explain the origin of the apparent mismatch between supergravity and CFT for extremal correlators involving giant gravitons in type IIB supergravity. We propose an ansatz for the wavefunctions of half-BPS giants which reproduces large $N$ limit of certain extremal two and three point functions in $\\mathcal{N}=4$ SYM.",
      "url": "http://arxiv.org/abs/2512.24979v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24979v1",
      "citations": null,
      "categories": [
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 496
    },
    {
      "title": "Attribution-Guided Distillation of Matryoshka Sparse Autoencoders",
      "authors": [
        "Cristina P. Martin-Linares",
        "Jonathan P. Ling"
      ],
      "abstract": "Sparse autoencoders (SAEs) aim to disentangle model activations into monosemantic, human-interpretable features. In practice, learned features are often redundant and vary across training runs and sparsity levels, which makes interpretations difficult to transfer and reuse. We introduce Distilled Matryoshka Sparse Autoencoders (DMSAEs), a training pipeline that distills a compact core of consistently useful features and reuses it to train new SAEs. DMSAEs run an iterative distillation cycle: train a Matryoshka SAE with a shared core, use gradient X activation to measure each feature's contribution to next-token loss in the most nested reconstruction, and keep only the smallest subset that explains a fixed fraction of the attribution. Only the core encoder weight vectors are transferred across cycles; the core decoder and all non-core latents are reinitialized each time. On Gemma-2-2B layer 12 residual stream activations, seven cycles of distillation (500M tokens, 65k width) yielded a distilled core of 197 features that were repeatedly selected. Training using this distilled core improves several SAEBench metrics and demonstrates that consistent sets of latent features can be transferred across sparsity levels",
      "url": "http://arxiv.org/abs/2512.24975v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24975v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 497
    },
    {
      "title": "Evaluating the Impact of Compression Techniques on the Robustness of CNNs under Natural Corruptions",
      "authors": [
        "Itallo Patrick Castro Alves Da Silva",
        "Emanuel Adler Medeiros Pereira",
        "Erick de Andrade Barboza",
        "Baldoino Fonseca dos Santos Neto",
        "Marcio de Medeiros Ribeiro"
      ],
      "abstract": "Compressed deep learning models are crucial for deploying computer vision systems on resource-constrained devices. However, model compression may affect robustness, especially under natural corruption. Therefore, it is important to consider robustness evaluation while validating computer vision systems. This paper presents a comprehensive evaluation of compression techniques - quantization, pruning, and weight clustering applied individually and in combination to convolutional neural networks (ResNet-50, VGG-19, and MobileNetV2). Using the CIFAR-10-C and CIFAR 100-C datasets, we analyze the trade-offs between robustness, accuracy, and compression ratio. Our results show that certain compression strategies not only preserve but can also improve robustness, particularly on networks with more complex architectures. Utilizing multiobjective assessment, we determine the best configurations, showing that customized technique combinations produce beneficial multi-objective results. This study provides insights into selecting compression methods for robust and efficient deployment of models in corrupted real-world environments.",
      "url": "http://arxiv.org/abs/2512.24971v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24971v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 498
    },
    {
      "title": "Large language models and the entropy of English",
      "authors": [
        "Colin Scheibner",
        "Lindsay M. Smith",
        "William Bialek"
      ],
      "abstract": "We use large language models (LLMs) to uncover long-ranged structure in English texts from a variety of sources. The conditional entropy or code length in many cases continues to decrease with context length at least to $N\\sim 10^4$ characters, implying that there are direct dependencies or interactions across these distances. A corollary is that there are small but significant correlations between characters at these separations, as we show from the data independent of models. The distribution of code lengths reveals an emergent certainty about an increasing fraction of characters at large $N$. Over the course of model training, we observe different dynamics at long and short context lengths, suggesting that long-ranged structure is learned only gradually. Our results constrain efforts to build statistical physics models of LLMs or language itself.",
      "url": "http://arxiv.org/abs/2512.24969v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24969v1",
      "citations": null,
      "categories": [
        "cond-mat.stat-mech",
        "cs.CL",
        "physics.bio-ph",
        "q-bio.NC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 499
    },
    {
      "title": "The least prime with a given cycle type",
      "authors": [
        "Peter J. Cho",
        "Robert J. Lemke Oliver",
        "Asif Zaman"
      ],
      "abstract": "Let $G$ be a finite group. Let $K/k$ be a Galois extension of number fields with Galois group isomorphic to $G$, and let $C \\subseteq \\mathrm{Gal}(K/k) \\simeq G$ be a conjugacy invariant subset. It is well known that there exists an unramified prime ideal $\\mathfrak{p}$ of $k$ with Frobenius element lying in $C$ and norm satisfying $\\mathrm{N}\\mathfrak{p} \\ll |\\mathrm{Disc}(K)|^α$ for some constant $α= α(G,C)$. There is a rich literature establishing unconditional admissible values for $α$, with most approaches proceeding by studying the zeros of $L$-functions. We give an alternative approach, not relying on zeros, that often substantially improves this exponent $α$ for any fixed finite group $G$, provided $C$ is a union of rational equivalence classes. As a particularly striking example, we prove that there exist absolute constants $c_1,c_2 > 0$ such that for any $n\\geq 2$ and any conjugacy class $C \\subset S_n$, one may take $α(S_n,C) = c_1 \\exp(-c_2n)$. Our approach reduces the core problem to a question in character theory.",
      "url": "http://arxiv.org/abs/2512.24963v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24963v1",
      "citations": null,
      "categories": [
        "math.NT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 500
    },
    {
      "title": "Fundamental Limits for Near-Field Sensing -- Part I: Narrow-Band Systems",
      "authors": [
        "Tong Wei",
        "Kumar Vijay Mishra",
        "Bhavani Shankar M. R.",
        "Björn Ottersten"
      ],
      "abstract": "Extremely large-scale antenna arrays (ELAAs) envisioned for 6G enable high-resolution sensing. However, the ELAAs worked in extremely high frequency will push operation into the near-field region, where spherical wavefronts invalidate classical far-field models and alter fundamental estimation limits. The purpose of this and the companion paper (Part II) is to develop the theory of fundamental limits for near-field sensing systems in detail. In this paper (Part I), we develop a unified narrow-band near-field signal model for joint parameter sensing of moving targets using the ELAAs. Leveraging the Slepian--Bangs formulation, we derive closed-form Cram'er--Rao bounds (CRBs) for joint estimation of target position, velocity, and radar cross-section (RCS) under the slow-time sampling model. To obtain interpretable insights, we further establish explicit far-field and near-field approximations that reveal how the bounds scale with array aperture, target range, carrier wavelength, and coherent integration length. The resulting expressions expose the roles of self-information terms and their cross terms, clarifying when Fresnel corrections become non-negligible and providing beamformer and algorithm design guidelines for near-field sensing with ELAAs. Simulation results validate the derived CRBs and their far-field and near-field approximations, demonstrating accurate agreement with the analytical scaling laws across representative array sizes and target ranges.",
      "url": "http://arxiv.org/abs/2512.24958v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24958v1",
      "citations": null,
      "categories": [
        "eess.SP"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 501
    },
    {
      "title": "Simulations of two-dimensional single-mode Rayleigh-Taylor Instability using front-tracking/ghost-fluid method: comparison to experiments and theory",
      "authors": [
        "James Burton",
        "Tulin Kaman"
      ],
      "abstract": "Two-dimensional single-mode Rayleigh-Taylor Instability (RTI) is simulated using an accurate and robust front-tracking/ghost-fluid method (FT/GFM) with high-order weighted essentially non-oscillatory (WENO) scheme. We compare our numerical results with the single-mode RTI experiments of Renoult, Rosenblatt and Carles (2015). The time evolution of the interface between two immiscible fluids and the effects of surface tension on the growth of the amplitude and asymmetry of the perturbed interface are examined for the initial wavelength 1 cm and the Atwood number A=0.29. The important features of RTI flows such as interface profiles, bubble/spike penetration and velocities show good agreement between experiments and simulations of immiscible fluids with surface tension. The velocity vector fields for the bubble and spike in the linear and nonlinear regimes are consistent with the theory for the single wavelength perturbation.",
      "url": "http://arxiv.org/abs/2512.24949v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24949v1",
      "citations": null,
      "categories": [
        "physics.flu-dyn"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 502
    },
    {
      "title": "RAIR: A Rule-Aware Benchmark Uniting Challenging Long-Tail and Visual Salience Subset for E-commerce Relevance Assessment",
      "authors": [
        "Chenji Lu",
        "Zhuo Chen",
        "Hui Zhao",
        "Zhenyi Wang",
        "Pengjie Wang",
        "Jian Xu",
        "Bo Zheng"
      ],
      "abstract": "Search relevance plays a central role in web e-commerce. While large language models (LLMs) have shown significant results on relevance task, existing benchmarks lack sufficient complexity for comprehensive model assessment, resulting in an absence of standardized relevance evaluation metrics across the industry. To address this limitation, we propose Rule-Aware benchmark with Image for Relevance assessment(RAIR), a Chinese dataset derived from real-world scenarios. RAIR established a standardized framework for relevance assessment and provides a set of universal rules, which forms the foundation for standardized evaluation. Additionally, RAIR analyzes essential capabilities required for current relevance models and introduces a comprehensive dataset consists of three subset: (1) a general subset with industry-balanced sampling to evaluate fundamental model competencies; (2) a long-tail hard subset focus on challenging cases to assess performance limits; (3) a visual salience subset for evaluating multimodal understanding capabilities. We conducted experiments on RAIR using 14 open and closed-source models. The results demonstrate that RAIR presents sufficient challenges even for GPT-5, which achieved the best performance. RAIR data are now available, serving as an industry benchmark for relevance assessment while providing new insights into general LLM and Visual Language Model(VLM) evaluation.",
      "url": "http://arxiv.org/abs/2512.24943v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24943v1",
      "citations": null,
      "categories": [
        "cs.IR",
        "cs.AI",
        "cs.CL",
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 503
    },
    {
      "title": "Iterative Deployment Improves Planning Skills in LLMs",
      "authors": [
        "Augusto B. Corrêa",
        "Yoav Gelberg",
        "Luckeciano C. Melo",
        "Ilia Shumailov",
        "André G. Pereira",
        "Yarin Gal"
      ],
      "abstract": "We show that iterative deployment of large language models (LLMs), each fine-tuned on data carefully curated by users from the previous models' deployment, can significantly change the properties of the resultant models. By testing this mechanism on various planning domains, we observe substantial improvements in planning skills, with later models displaying emergent generalization by discovering much longer plans than the initial models. We then provide theoretical analysis showing that iterative deployment effectively implements reinforcement learning (RL) training in the outer-loop (i.e. not as part of intentional model training), with an implicit reward function. The connection to RL has two important implications: first, for the field of AI safety, as the reward function entailed by repeated deployment is not defined explicitly, and could have unexpected implications to the properties of future model deployments. Second, the mechanism highlighted here can be viewed as an alternative training regime to explicit RL, relying on data curation rather than explicit rewards.",
      "url": "http://arxiv.org/abs/2512.24940v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24940v1",
      "citations": null,
      "categories": [
        "cs.AI",
        "cs.CL",
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 504
    },
    {
      "title": "Vibe Coding, Interface Flattening",
      "authors": [
        "Hongrui Jin"
      ],
      "abstract": "Large language models are reshaping programming by enabling 'vibe coding': the development of softwares through natural-language interaction with model-driven toolchains. This article argues that vibe coding is best understood as interface flattening, a reconfiguration in which previously distinct modalities (GUI, CLI, and API) appear to converge into a single conversational surface, even as the underlying chain of translation from intention to machinic effect lengthens and thickens. Drawing on Friedrich Kittler's materialist media theory and Alexander Galloway's account of interfaces as sites of protocol control, the paper situates programming as a historically localised interface arrangement rather than an essential relation to computation. Through a materialist reconstruction of the contemporary vibe-coding stack, it shows how remote compute infrastructures, latency and connectivity, structured outputs, function/tool calling, and interoperability standards such as the Model Context Protocol relocate control and meaning-making power to model and protocol providers. The apparent democratisation of technical capability therefore depends on new dependencies and new literacies. By foregrounding the tension between experiential flattening and infrastructural thickening, I demonstrate how LLM-mediated development redistributes symbolic labour/power, obscures responsibility, and privatises competencies previously dispersed across programming communities, contributing a critical lens on the political economy of AI-mediated human-computer interaction.",
      "url": "http://arxiv.org/abs/2512.24939v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24939v1",
      "citations": null,
      "categories": [
        "cs.HC",
        "cs.CL"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 505
    },
    {
      "title": "Modelling the movements of organisms by stochastic theory in a comoving frame",
      "authors": [
        "Norberto Lucero Azuara",
        "Rainer Klages"
      ],
      "abstract": "Imagine you walk in a plane. You move by making a step of a certain length per time interval in a chosen direction. Repeating this process by randomly sampling step length and turning angle defines a two-dimensional random walk in what we call comoving frame coordinates. This is precisely how Ross and Pearson proposed to model the movements of organisms more than a century ago. Decades later their concept was generalised by including persistence leading to a correlated random walk, which became a popular model in Movement Ecology. In contrast, Langevin equations describing cell migration and used in active matter theory are typically formulated by position and velocity in a fixed Cartesian frame. In this article, we explore the transformation of stochastic Langevin dynamics from the Cartesian into the comoving frame. We show that the Ornstein-Uhlenbeck process for the Cartesian velocity of a walker can be transformed exactly into a stochastic process that is defined self-consistently in the comoving frame, thereby profoundly generalising correlated random walk models. This approach yields a general conceptual framework how to transform stochastic processes from the Cartesian into the comoving frame. Our theory paves the way to derive, invent and explore novel stochastic processes in the comoving frame for modelling the movements of organisms. It can also be applied to design novel stochastic dynamics for autonomously moving robots and drones.",
      "url": "http://arxiv.org/abs/2512.24937v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24937v1",
      "citations": null,
      "categories": [
        "physics.bio-ph",
        "math-ph",
        "math.DS"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 506
    },
    {
      "title": "Green's function on the Tate curve",
      "authors": [
        "An Huang",
        "Rebecca Rohrlich",
        "Yaojia Sun",
        "Eric Whyman"
      ],
      "abstract": "Motivated by the question of defining a $p$-adic string worldsheet action in genus one, we define a Laplacian operator on the Tate curve, and study its Green's function. We show that the Green's function exists. We provide an explicit formula for the Green's function, which turns out to be a non-Archimedean counterpart of the Archimedean Green's function on a flat torus.",
      "url": "http://arxiv.org/abs/2512.24935v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24935v1",
      "citations": null,
      "categories": [
        "math.NT",
        "hep-th",
        "math-ph",
        "math.AP"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 507
    },
    {
      "title": "Adaptive Dependency-aware Prompt Optimization Framework for Multi-Step LLM Pipeline",
      "authors": [
        "Minjun Zhao",
        "Xinyu Zhang",
        "Shuai Zhang",
        "Deyang Li",
        "Ruifeng Shi"
      ],
      "abstract": "Multi-step LLM pipelines invoke large language models multiple times in a structured sequence and can effectively solve complex tasks, but their performance heavily depends on the prompts used at each step. Jointly optimizing these prompts is difficult due to missing step-level supervision and inter-step dependencies. Existing end-to-end prompt optimization methods struggle under these conditions and often yield suboptimal or unstable updates. We propose ADOPT, an Adaptive Dependency-aware Prompt Optimization framework for multi-step LLM pipelines. ADOPT explicitly models the dependency between each LLM step and the final task outcome, enabling precise text-gradient estimation analogous to computing analytical derivatives. It decouples textual gradient estimation from gradient updates, reducing multi-prompt optimization to flexible single-prompt optimization steps, and employs a Shapley-based mechanism to adaptively allocate optimization resources. Experiments on real-world datasets and diverse pipeline structures show that ADOPT is effective and robust, consistently outperforming state-of-the-art prompt optimization baselines.",
      "url": "http://arxiv.org/abs/2512.24933v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24933v1",
      "citations": null,
      "categories": [
        "cs.CL",
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 508
    },
    {
      "title": "Generalised Hermite-Einstein Fibre Metrics and Slope Stability for Holomorphic Vector Bundles",
      "authors": [
        "Dan Popovici"
      ],
      "abstract": "Let $X$ be a compact complex manifold of dimension $n$ and let $m$ be a positive integer with $m\\leq n$. Assume that $X$ admits a Kähler metric $ω$ and a weakly positive, $\\partial\\bar\\partial$-closed, smooth $(n-m,\\,n-m)$-form $Ω$. We introduce the notions of $(ω,\\,Ω)$-Hermite-Einstein holomorphic vector bundles and $(ω,\\,Ω)$(-semi)-stable coherent sheaves on $X$ by generalising the classical definitions depending only on $ω$. We then prove that the $(ω,\\,Ω)$-Hermite-Einstein condition implies the $(ω,\\,Ω)$-semi-stability of a holomorphic vector bundle and its splitting into $(ω,\\,Ω)$-stable subbundles. This extends a classical result by Kobayashi and Lübke to our generalised setting.",
      "url": "http://arxiv.org/abs/2512.24932v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24932v1",
      "citations": null,
      "categories": [
        "math.AG",
        "math.CV",
        "math.DG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 509
    },
    {
      "title": "Introduction to black hole thermodynamics",
      "authors": [
        "Pietro Benetti Genolini"
      ],
      "abstract": "These are the lecture notes for a course at the \"Roberto Salmeron School in Mathematical Physics\" held at the University of Brasilia in September 2025, to be published in the proceedings book \"Modern topics in mathematical physics.\" The course provides a concise and biased introduction to black hole thermodynamics. It covers the laws of black hole mechanics, Hawking radiation, Euclidean quantum gravity methods, and AdS black holes.",
      "url": "http://arxiv.org/abs/2512.24929v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24929v1",
      "citations": null,
      "categories": [
        "hep-th",
        "gr-qc"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 510
    },
    {
      "title": "Are First-Order Diffusion Samplers Really Slower? A Fast Forward-Value Approach",
      "authors": [
        "Yuchen Jiao",
        "Na Li",
        "Changxiao Cai",
        "Gen Li"
      ],
      "abstract": "Higher-order ODE solvers have become a standard tool for accelerating diffusion probabilistic model (DPM) sampling, motivating the widespread view that first-order methods are inherently slower and that increasing discretization order is the primary path to faster generation. This paper challenges this belief and revisits acceleration from a complementary angle: beyond solver order, the placement of DPM evaluations along the reverse-time dynamics can substantially affect sampling accuracy in the low-neural function evaluation (NFE) regime.\n  We propose a novel training-free, first-order sampler whose leading discretization error has the opposite sign to that of DDIM. Algorithmically, the method approximates the forward-value evaluation via a cheap one-step lookahead predictor. We provide theoretical guarantees showing that the resulting sampler provably approximates the ideal forward-value trajectory while retaining first-order convergence. Empirically, across standard image generation benchmarks (CIFAR-10, ImageNet, FFHQ, and LSUN), the proposed sampler consistently improves sample quality under the same NFE budget and can be competitive with, and sometimes outperform, state-of-the-art higher-order samplers. Overall, the results suggest that the placement of DPM evaluations provides an additional and largely independent design angle for accelerating diffusion sampling.",
      "url": "http://arxiv.org/abs/2512.24927v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24927v1",
      "citations": null,
      "categories": [
        "stat.ML",
        "cs.LG",
        "math.ST"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 511
    },
    {
      "title": "Semi-Supervised Diversity-Aware Domain Adaptation for 3D Object detection",
      "authors": [
        "Bartłomiej Olber",
        "Jakub Winter",
        "Paweł Wawrzyński",
        "Andrii Gamalii",
        "Daniel Górniak",
        "Marcin Łojek",
        "Robert Nowak",
        "Krystian Radlak"
      ],
      "abstract": "3D object detectors are fundamental components of perception systems in autonomous vehicles. While these detectors achieve remarkable performance on standard autonomous driving benchmarks, they often struggle to generalize across different domains - for instance, a model trained in the U.S. may perform poorly in regions like Asia or Europe. This paper presents a novel lidar domain adaptation method based on neuron activation patterns, demonstrating that state-of-the-art performance can be achieved by annotating only a small, representative, and diverse subset of samples from the target domain if they are correctly selected. The proposed approach requires very small annotation budget and, when combined with post-training techniques inspired by continual learning prevent weight drift from the original model. Empirical evaluation shows that the proposed domain adaptation approach outperforms both linear probing and state-of-the-art domain adaptation techniques.",
      "url": "http://arxiv.org/abs/2512.24922v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24922v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 512
    },
    {
      "title": "Valence quark distribution of the pion inside a medium with finite baryon density: A Nambu--Jona-Lasinio model approach",
      "authors": [
        "Ashutosh Dwibedi",
        "Satyajit Puhan",
        "Sabyasachi Ghosh",
        "Harleen Dahiya"
      ],
      "abstract": "We calculate the in-medium valence quark distribution of the pion immersed in a finite baryon density using the light-cone quark model. The medium-modified pion properties are obtained by using the constituent quark mass-dependent light cone wave functions. To obtain the constituent quark masses at finite baryon density, we employ the two-flavor Nambu--Jona-Lasinio model. We primarily focus on the in-medium electromagnetic form factor, distribution amplitude, and the parton distribution function of the pion. The parton distribution functions are also evolved from the model scale to a perturbative scale using next to leading order Dokshitzer-Gribov-Lipatov-Altarelli-Parisi evolution equations. Furthermore, our calculated form factors are compared with available experimental measurements and lattice quantum chromodynamics studies. We also examine the Mellin moments derived from our parton distribution functions in comparison with existing extractions and theoretical model predictions.",
      "url": "http://arxiv.org/abs/2512.24921v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24921v1",
      "citations": null,
      "categories": [
        "hep-ph",
        "nucl-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 513
    },
    {
      "title": "Transgression in the primitive cohomology",
      "authors": [
        "Hao Zhuang"
      ],
      "abstract": "We study the Chern-Weil theory for the primitive cohomology of a symplectic manifold. First, given a symplectic manifold, we review the superbundle-valued forms on this manifold and prove a primitive version of the Bianchi identity. Second, as the main result, we prove a transgression formula associated with the boundary map of the primitive cohomology. Third, as an application of the main result, we introduce the concept of primitive characteristic classes and point out a further direction.",
      "url": "http://arxiv.org/abs/2512.24920v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24920v1",
      "citations": null,
      "categories": [
        "math.DG",
        "math-ph",
        "math.SG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 514
    },
    {
      "title": "Property (T) and Poincaré duality in dimension three",
      "authors": [
        "Cameron Gates Rudd"
      ],
      "abstract": "We use a recent result of Bader and Sauer on coboundary expansion to prove residually finite three-dimensional Poincaré duality groups never have property (T). This implies such groups are never Kähler. The argument applies to fundamental groups of (possibly non-aspherical) compact 3-manifolds, giving a new proof of a theorem of Fujiwara that states if the fundamental group of a compact 3-manifold has property (T), then that group is finite. The only consequence of geometrization needed in the proof is that 3-manifold groups are residually finite.",
      "url": "http://arxiv.org/abs/2512.24919v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24919v1",
      "citations": null,
      "categories": [
        "math.GT",
        "math.GR"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 515
    },
    {
      "title": "Frequent subgraph-based persistent homology for graph classification",
      "authors": [
        "Xinyang Chen",
        "Amaël Broustet",
        "Guoting Chen"
      ],
      "abstract": "Persistent homology (PH) has recently emerged as a powerful tool for extracting topological features. Integrating PH into machine learning and deep learning models enhances topology awareness and interpretability. However, most PH methods on graphs rely on a limited set of filtrations, such as degree-based or weight-based filtrations, which overlook richer features like recurring information across the dataset and thus restrict expressive power. In this work, we propose a novel graph filtration called Frequent Subgraph Filtration (FSF), which is derived from frequent subgraphs and produces stable and information-rich frequency-based persistent homology (FPH) features. We study the theoretical properties of FSF and provide both proofs and experimental validation. Beyond persistent homology itself, we introduce two approaches for graph classification: an FPH-based machine learning model (FPH-ML) and a hybrid framework that integrates FPH with graph neural networks (FPH-GNNs) to enhance topology-aware graph representation learning. Our frameworks bridge frequent subgraph mining and topological data analysis, offering a new perspective on topology-aware feature extraction. Experimental results show that FPH-ML achieves competitive or superior accuracy compared with kernel-based and degree-based filtration methods. When integrated into graph neural networks, FPH yields relative performance gains ranging from 0.4 to 21 percent, with improvements of up to 8.2 percentage points over GCN and GIN backbones across benchmarks.",
      "url": "http://arxiv.org/abs/2512.24917v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24917v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "math.AT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 516
    },
    {
      "title": "AI-Driven Cloud Resource Optimization for Multi-Cluster Environments",
      "authors": [
        "Vinoth Punniyamoorthy",
        "Akash Kumar Agarwal",
        "Bikesh Kumar",
        "Abhirup Mazumder",
        "Kabilan Kannan",
        "Sumit Saha"
      ],
      "abstract": "Modern cloud-native systems increasingly rely on multi-cluster deployments to support scalability, resilience, and geographic distribution. However, existing resource management approaches remain largely reactive and cluster-centric, limiting their ability to optimize system-wide behavior under dynamic workloads. These limitations result in inefficient resource utilization, delayed adaptation, and increased operational overhead across distributed environments. This paper presents an AI-driven framework for adaptive resource optimization in multi-cluster cloud systems. The proposed approach integrates predictive learning, policy-aware decision-making, and continuous feedback to enable proactive and coordinated resource management across clusters. By analyzing cross-cluster telemetry and historical execution patterns, the framework dynamically adjusts resource allocation to balance performance, cost, and reliability objectives. A prototype implementation demonstrates improved resource efficiency, faster stabilization during workload fluctuations, and reduced performance variability compared to conventional reactive approaches. The results highlight the effectiveness of intelligent, self-adaptive infrastructure management as a key enabler for scalable and resilient cloud platforms.",
      "url": "http://arxiv.org/abs/2512.24914v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24914v1",
      "citations": null,
      "categories": [
        "cs.DC",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 517
    },
    {
      "title": "On Diophantine exponents of lattices",
      "authors": [
        "Nikolay Moshchevitin"
      ],
      "abstract": "We describe the spectrum of ordinary Diophantine exponents for $d$-dimensional lattices. The result reduces the problem to two-dimensional case and uses argument of metric theory.",
      "url": "http://arxiv.org/abs/2512.24913v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24913v1",
      "citations": null,
      "categories": [
        "math.NT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 518
    },
    {
      "title": "Non-Equilibrium Dynamics in QCD and Holography",
      "authors": [
        "Matthias Kaminski"
      ],
      "abstract": "The plasma generated in heavy ion collisions goes through different phases in its time evolution. While early times right after the collision are governed by far-from equilibrium dynamics, later times are believed to be well described by near-equilibrium dynamics. While the regimes of non-equilibrium are prohibitively complicated to describe within QCD, effective descriptions such as hydrodynamics provide a viable approach. In addition, holographic descriptions allow access to the full non-equilibrium dynamics at strong coupling. In this presentation, we review three examples of such hydrodynamic approaches and corresponding holographic descriptions: 1) non-equilibrium shear viscosity, 2) propagation of non-equilibrium sound waves, and 3) the non-equilibrium chiral magnetic effect.",
      "url": "http://arxiv.org/abs/2512.24909v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24909v1",
      "citations": null,
      "categories": [
        "nucl-th",
        "hep-th",
        "nucl-ex"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 519
    },
    {
      "title": "Spectral Graph Neural Networks for Cognitive Task Classification in fMRI Connectomes",
      "authors": [
        "Debasis Maji",
        "Arghya Banerjee",
        "Debaditya Barman"
      ],
      "abstract": "Cognitive task classification using machine learning plays a central role in decoding brain states from neuroimaging data. By integrating machine learning with brain network analysis, complex connectivity patterns can be extracted from functional magnetic resonance imaging connectomes. This process transforms raw blood-oxygen-level-dependent (BOLD) signals into interpretable representations of cognitive processes. Graph neural networks (GNNs) further advance this paradigm by modeling brain regions as nodes and functional connections as edges, capturing topological dependencies and multi-scale interactions that are often missed by conventional approaches. Our proposed SpectralBrainGNN model, a spectral convolution framework based on graph Fourier transforms (GFT) computed via normalized Laplacian eigendecomposition. Experiments on the Human Connectome Project-Task (HCPTask) dataset demonstrate the effectiveness of the proposed approach, achieving a classification accuracy of 96.25\\%. The implementation is publicly available at https://github.com/gnnplayground/SpectralBrainGNN to support reproducibility and future research.",
      "url": "http://arxiv.org/abs/2512.24901v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24901v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 520
    },
    {
      "title": "PRISM: A hierarchical multiscale approach for time series forecasting",
      "authors": [
        "Zihao Chen",
        "Alexandre Andre",
        "Wenrui Ma",
        "Ian Knight",
        "Sergey Shuvaev",
        "Eva Dyer"
      ],
      "abstract": "Forecasting is critical in areas such as finance, biology, and healthcare. Despite the progress in the field, making accurate forecasts remains challenging because real-world time series contain both global trends, local fine-grained structure, and features on multiple scales in between. Here, we present a new forecasting method, PRISM (Partitioned Representation for Iterative Sequence Modeling), that addresses this challenge through a learnable tree-based partitioning of the signal. At the root of the tree, a global representation captures coarse trends in the signal, while recursive splits reveal increasingly localized views of the signal. At each level of the tree, data are projected onto a time-frequency basis (e.g., wavelets or exponential moving averages) to extract scale-specific features, which are then aggregated across the hierarchy. This design allows the model to jointly capture global structure and local dynamics of the signal, enabling accurate forecasting. Experiments across benchmark datasets show that our method outperforms state-of-the-art methods for forecasting. Overall, these results demonstrate that our hierarchical approach provides a lightweight and flexible framework for forecasting multivariate time series. The code is available at https://github.com/nerdslab/prism.",
      "url": "http://arxiv.org/abs/2512.24898v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24898v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 521
    },
    {
      "title": "Self-Supervised Amortized Neural Operators for Optimal Control: Scaling Laws and Applications",
      "authors": [
        "Wuzhe Xu",
        "Jiequn Han",
        "Rongjie Lai"
      ],
      "abstract": "Optimal control provides a principled framework for transforming dynamical system models into intelligent decision-making, yet classical computational approaches are often too expensive for real-time deployment in dynamic or uncertain environments. In this work, we propose a method based on self-supervised neural operators for open-loop optimal control problems. It offers a new paradigm by directly approximating the mapping from system conditions to optimal control strategies, enabling instantaneous inference across diverse scenarios once trained. We further extend this framework to more complex settings, including dynamic or partially observed environments, by integrating the learned solution operator with Model Predictive Control (MPC). This yields a solution-operator learning method for closed-loop control, in which the learned operator supplies rapid predictions that replace the potentially time-consuming optimization step in conventional MPC. This acceleration comes with a quantifiable price to pay. Theoretically, we derive scaling laws that relate generalization error and sample/model complexity to the intrinsic dimension of the problem and the regularity of the optimal control function. Numerically, case studies show efficient, accurate real-time performance in low-intrinsic-dimension regimes, while accuracy degrades as problem complexity increases. Together, these results provide a balanced perspective: neural operators are a powerful novel tool for high-performance control when hidden low-dimensional structure can be exploited, yet they remain fundamentally constrained by the intrinsic dimensional complexity in more challenging settings.",
      "url": "http://arxiv.org/abs/2512.24897v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24897v1",
      "citations": null,
      "categories": [
        "math.OC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 522
    },
    {
      "title": "Interior structure of black holes with nonlinear terms",
      "authors": [
        "Zi-Qiang Zhao",
        "Zhang-Yu Nie",
        "Xing-Kun Zhang",
        "Yu-Sen An",
        "Jing-Fei Zhang",
        "Xin Zhang"
      ],
      "abstract": "We investigate the oscillation of the Kasner exponent $p_t$ near critical point of the hairy black holes dual to holographic superfluid and reveal a clear inverse periodicity $f(T_c/(T_c-T))$ in a large region below the critical temperature. We first introduce the fourth-power term with a coefficient $λ$ to adjust the oscillatory behavior of the Kasner exponent $p_t$ near the critical point. Importantly, we show that the nonlinear coefficient $λ$ provides accurate control of this periodicity: a positive $λ$ stretches the region, while a negative $λ$ compresses it. By contrast, the influence of another coefficient $τ$ is more concentrated in regions away from the critical point. This work provides a new perspective for understanding the complex dynamical structure inside black holes and extends the actively control from the fourth- and sixth-power term into the black hole interior region.",
      "url": "http://arxiv.org/abs/2512.24893v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24893v1",
      "citations": null,
      "categories": [
        "gr-qc",
        "hep-ph",
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 523
    },
    {
      "title": "Bubbling wormholes and matrix models",
      "authors": [
        "Panos Betzios",
        "Ji Hoon Lee",
        "Olga Papadoulaki",
        "Yanjun Zhou"
      ],
      "abstract": "The thermofield double state entangles two copies of a CFT via a sum over energy eigenstates and is dual to the two-sided eternal black hole. We explore an analogous construction using sums over gauge group representations of half-BPS Wilson loops in multiple copies of $U(N)$ $\\mathcal{N}=4$ super Yang-Mills. These sums act as delta function-like operators that correlate the eigenvalues of the corresponding half-BPS matrix models. We suggest that the holographic duals are ''bubbling wormhole'' geometries: multi-covers of AdS$_5$ $\\times S^5$ whose conformal boundary consists of multiple four-spheres intersecting on a common circle. We analyze the matrix model free energy, discuss its bulk interpretation, and study probe loops in these backgrounds.",
      "url": "http://arxiv.org/abs/2512.24891v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24891v1",
      "citations": null,
      "categories": [
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 524
    },
    {
      "title": "Coherent span-valued 2D TQFTs",
      "authors": [
        "Sophia E Marx",
        "Rajan Amit Mehta"
      ],
      "abstract": "We consider commutative Frobenius pseudomonoids in the bicategory of spans, and we show that they are in correspondence with 2-Segal cosymmetric sets. Such a structure can be interpreted as a coherent 2-dimensional topological quantum field theory taking values in the bicategory of spans. We also describe a construction that produces a 2-Segal cosymmetric set from any partial monoid equipped with a distinguished element.",
      "url": "http://arxiv.org/abs/2512.24887v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24887v1",
      "citations": null,
      "categories": [
        "math.AT",
        "math.CT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 525
    },
    {
      "title": "Heterogeneous Multi-Agent Multi-Target Tracking using Cellular Sheaves",
      "authors": [
        "Tyler Hanks",
        "Cristian F. Nino",
        "Joana Bou Barcelo",
        "Austin Copeland",
        "Warren Dixon",
        "James Fairbanks"
      ],
      "abstract": "Multi-agent target tracking in the presence of nonlinear dynamics and agent heterogeneity, where state-space dimensions may differ, is a challenging problem that traditional graph Laplacian methods cannot easily address. This work leverages the framework of cellular sheaves, a mathematical generalization of graph theory, to natively model such heterogeneous systems. While existing coordination sheaf frameworks focus on cooperative problems like consensus, this work extends them to the non-cooperative target-tracking problem. The tracking of multiple, unknown targets is formulated as a harmonic extension problem on a cellular sheaf, accommodating nonlinear dynamics and external disturbances for all agents. A decentralized control law is developed using the sheaf Laplacian, and a corresponding Lyapunov-based stability analysis is provided to guarantee tracking error convergence, with results validated by simulation.",
      "url": "http://arxiv.org/abs/2512.24886v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24886v1",
      "citations": null,
      "categories": [
        "eess.SY",
        "cs.MA",
        "math.AT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 526
    },
    {
      "title": "BEDA: Belief Estimation as Probabilistic Constraints for Performing Strategic Dialogue Acts",
      "authors": [
        "Hengli Li",
        "Zhaoxin Yu",
        "Qi Shen",
        "Chenxi Li",
        "Mengmeng Wang",
        "Tinglang Wu",
        "Yipeng Kang",
        "Yuxuan Wang",
        "Song-Chun Zhu",
        "Zixia Jia"
      ],
      "abstract": "Strategic dialogue requires agents to execute distinct dialogue acts, for which belief estimation is essential. While prior work often estimates beliefs accurately, it lacks a principled mechanism to use those beliefs during generation. We bridge this gap by first formalizing two core acts Adversarial and Alignment, and by operationalizing them via probabilistic constraints on what an agent may generate. We instantiate this idea in BEDA, a framework that consists of the world set, the belief estimator for belief estimation, and the conditional generator that selects acts and realizes utterances consistent with the inferred beliefs. Across three settings, Conditional Keeper Burglar (CKBG, adversarial), Mutual Friends (MF, cooperative), and CaSiNo (negotiation), BEDA consistently outperforms strong baselines: on CKBG it improves success rate by at least 5.0 points across backbones and by 20.6 points with GPT-4.1-nano; on Mutual Friends it achieves an average improvement of 9.3 points; and on CaSiNo it achieves the optimal deal relative to all baselines. These results indicate that casting belief estimation as constraints provides a simple, general mechanism for reliable strategic dialogue.",
      "url": "http://arxiv.org/abs/2512.24885v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24885v1",
      "citations": null,
      "categories": [
        "cs.CL",
        "cs.GT",
        "cs.MA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 527
    },
    {
      "title": "Probing quantum-coherent dynamics with free electrons",
      "authors": [
        "H. B. Crispin",
        "N. Talebi"
      ],
      "abstract": "Recent advances in time-resolved cathodoluminescence have enabled ultrafast studies of single emitters in quantum materials with femtosecond temporal resolution. Here, we develop a quantum theory modeling the dynamics of free electrons interacting with quantum emitters in arbitrary initial states. Our analysis reveals that a free electron can induce transient coherent oscillations in the populations when the system is initially prepared in a coherent superposition of its states. Moreover, the electron energy spectrum exhibits a clear signature of the quantum coherence and sensitivity to the transition frequency of the emitter. These coherence effects manifest themselves as oscillations in the zero-loss peak of the spectral energy-loss probability. Our findings pave the way for characterization of quantum-coherent dynamics of individual quantum emitters by electron-probes.",
      "url": "http://arxiv.org/abs/2512.24883v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24883v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 528
    },
    {
      "title": "Description of Baryon Mass Spectrum by Open Strings and Diquarks",
      "authors": [
        "Yuki Fujimoto"
      ],
      "abstract": "We analyze the mass spectra of hadrons and demonstrate that the physical spectra of mesons and baryons are well described by the exponential spectrum of open strings. The open string spectrum, derived from string theory, is characterized by a unique Hagedorn temperature $T_{\\rm H}$ and free from any other parameters. Notably, our fitting to the physical spectra yields consistent values for both mesons and baryons, $T_{\\rm H} \\simeq 0.34\\,\\text{GeV}$, which contrasts with previous phenomenological analyses that suggested different values. This obtained value aligns well with typical string tension derived from lattice-QCD calculations and the Regge slope. In the baryonic sector, our results indicate that diquarks play a crucial role in describing the mass spectrum, implying that baryons can be understood as a quark-diquark system, as anticipated by Regge phenomenology. These findings have significant implications for our understanding of quark deconfinement, especially in the possibly existing regime at high temperature and small baryon chemical potential within the QCD phase diagram.",
      "url": "http://arxiv.org/abs/2512.24882v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24882v1",
      "citations": null,
      "categories": [
        "hep-ph",
        "nucl-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 529
    },
    {
      "title": "Exact Identity Linking Entropy Production and Mutual Information",
      "authors": [
        "Doohyeong Cho",
        "Hawoong Jeong"
      ],
      "abstract": "Linking entropy production (EP) to information is a key step toward data-driven nonequilibrium thermodynamics. We derive an exact identity for overdamped Langevin dynamics that equates the total EP rate to the mutual-information rate between an infinitesimal displacement and its time-symmetric midpoint, up to a bulk mean-flow contribution. This mapping elevates information theory to a thermodynamic calculus: the chain rule yields a canonical, nonnegative split into self and interaction EP, and leads to a tighter bound on learning rate with interaction EP as the necessary cost. As a proof of concept, applying the estimator to red-blood-cell flickering shows that interaction EP robustly exposes active signatures that conventional summaries can miss.",
      "url": "http://arxiv.org/abs/2512.24877v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24877v1",
      "citations": null,
      "categories": [
        "cond-mat.stat-mech",
        "physics.bio-ph",
        "physics.data-an"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 530
    },
    {
      "title": "Insights on the homogeneous $3$-local representations of the twin groups",
      "authors": [
        "Mohamad N. Nasser"
      ],
      "abstract": "We provide a complete classification of the homogeneous $3$-local representations of the twin group $T_n$, the virtual twin group $VT_n$, and the welded twin group $WT_n$, for all $n\\geq 4$. Beyond this classification, we examine the main characteristics of these representations, particularly their irreducibility and faithfulness. More deeply, we show that all such representations are reducible, and most of them are unfaithful. Also, we find necessary and sufficient conditions of the first two types of the classified representations of $T_n$ to be irreducible in the case $n=4$. The obtained results provide insights into the algebraic structure of these three groups.",
      "url": "http://arxiv.org/abs/2512.24874v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24874v1",
      "citations": null,
      "categories": [
        "math.RT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 531
    },
    {
      "title": "Let It Flow: Agentic Crafting on Rock and Roll, Building the ROME Model within an Open Agentic Learning Ecosystem",
      "authors": [
        "Weixun Wang",
        "XiaoXiao Xu",
        "Wanhe An",
        "Fangwen Dai",
        "Wei Gao",
        "Yancheng He",
        "Ju Huang",
        "Qiang Ji",
        "Hanqi Jin",
        "Xiaoyang Li"
      ],
      "abstract": "Agentic crafting requires LLMs to operate in real-world environments over multiple turns by taking actions, observing outcomes, and iteratively refining artifacts. Despite its importance, the open-source community lacks a principled, end-to-end ecosystem to streamline agent development. We introduce the Agentic Learning Ecosystem (ALE), a foundational infrastructure that optimizes the production pipeline for agent LLMs. ALE consists of three components: ROLL, a post-training framework for weight optimization; ROCK, a sandbox environment manager for trajectory generation; and iFlow CLI, an agent framework for efficient context engineering. We release ROME (ROME is Obviously an Agentic Model), an open-source agent grounded by ALE and trained on over one million trajectories. Our approach includes data composition protocols for synthesizing complex behaviors and a novel policy optimization algorithm, Interaction-based Policy Alignment (IPA), which assigns credit over semantic interaction chunks rather than individual tokens to improve long-horizon training stability. Empirically, we evaluate ROME within a structured setting and introduce Terminal Bench Pro, a benchmark with improved scale and contamination control. ROME demonstrates strong performance across benchmarks like SWE-bench Verified and Terminal Bench, proving the effectiveness of the ALE infrastructure.",
      "url": "http://arxiv.org/abs/2512.24873v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24873v1",
      "citations": null,
      "categories": [
        "cs.AI",
        "cs.CL"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 532
    },
    {
      "title": "Configuration Spaces of Finite Representation Type Algebras",
      "authors": [
        "Nima Arkani-Hamed",
        "Hadleigh Frost",
        "Pierre-Guy Plamondon",
        "Giulio Salvatori",
        "Hugh Thomas"
      ],
      "abstract": "To every finite-dimensional $\\mathbb C$-algebra $Λ$ of finite representation type we associate an affine variety. These varieties are a large generalization of the varieties defined by \"$u$ variables\" satisfying \"$u$-equations\", first introduced in the context of open string theory and moduli space of ordered points on the real projective line by Koba and Nielsen, rediscovered by Brown as \"dihedral co-ordinates\", and recently generalized to any finite type hereditary algebras. We show that each such variety is irreducible and admits a rational parametrization. The assignment is functorial: algebra quotients correspond to monomial maps among the varieties. The non-negative real part of each variety has boundary strata that are controlled by Jasso reduction. These non-negative parts naturally define a generalization of open string integrals in physics, exhibiting factorization and splitting properties that do not come from a worldsheet picture. We further establish a family of Rogers dilogarithm identities extending results of Chapoton beyond the Dynkin case.",
      "url": "http://arxiv.org/abs/2512.24870v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24870v1",
      "citations": null,
      "categories": [
        "math.RT",
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 533
    },
    {
      "title": "Characterization of Transfer Using Multi-task Learning Curves",
      "authors": [
        "András Millinghoffer",
        "Bence Bolgár",
        "Péter Antal"
      ],
      "abstract": "Transfer effects manifest themselves both during training using a fixed data set and in inductive inference using accumulating data. We hypothesize that perturbing the data set by including more samples, instead of perturbing the model by gradient updates, provides a complementary and more fundamental characterization of transfer effects. To capture this phenomenon, we quantitatively model transfer effects using multi-task learning curves approximating the inductive performance over varying sample sizes. We describe an efficient method to approximate multi-task learning curves analogous to the Task Affinity Grouping method applied during training. We compare the statistical and computational approaches to transfer, which indicates considerably higher compute costs for the previous but better power and broader applicability. Evaluations are performed using a benchmark drug-target interaction data set. Our results show that learning curves can better capture the effects of multi-task learning and their multi-task extensions can delineate pairwise and contextual transfer effects in foundation models.",
      "url": "http://arxiv.org/abs/2512.24866v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24866v1",
      "citations": null,
      "categories": [
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 534
    },
    {
      "title": "Approximate Computation via Le Cam Simulability",
      "authors": [
        "Deniz Akdemir"
      ],
      "abstract": "We propose a decision-theoretic framework for computational complexity, complementary to classical theory: moving from syntactic exactness (Turing / Shannon) to semantic simulability (Le Cam). While classical theory classifies problems by the cost of exact solution, modern computation often seeks only decision-valid approximations. We introduce a framework where \"computation\" is viewed as the efficient simulation of a target statistical experiment within a bounded risk distortion (Le Cam deficiency).\n  We formally define computational deficiency ($δ_{\\text{poly}}$) and use it to construct the complexity class LeCam-P (Decision-Robust Polynomial Time), characterizing problems that may be syntactically hard but semantically easy to approximate. We show that classical Karp reductions can be viewed as zero-deficiency simulations, and that approximate reductions correspond to bounded deficiency. Furthermore, we establish the No-Free-Transfer Inequality, showing that strictly invariant representations inevitably destroy decision-relevant information. This framework offers a statistical perspective on approximation theory, bridging the gap between algorithmic complexity and decision theory.",
      "url": "http://arxiv.org/abs/2512.24860v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24860v1",
      "citations": null,
      "categories": [
        "math.ST",
        "cs.CC",
        "cs.IT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 535
    },
    {
      "title": "On a conjecture of Almgren II: area-minimizing submanifolds with fractal singular sets on almost any manifold",
      "authors": [
        "Zhenhua Liu"
      ],
      "abstract": "This paper is the second in a two-part solution to Almgren's conjecture on the existence of area-minimizing submanifolds with fractal singular sets. In part one, we construct area-minimizing submanifolds with fractal singular sets on certain special manifolds. Here we continue our work and show that area-minimizing submanifolds with fractal singular sets exist on almost any smooth manifold.",
      "url": "http://arxiv.org/abs/2512.24859v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24859v1",
      "citations": null,
      "categories": [
        "math.DG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 536
    },
    {
      "title": "Measuring Mixed-State Topological Invariant in Open Photonic Quantum Walk",
      "authors": [
        "Qin-Qin Wang",
        "Xiao-Ye Xu",
        "Yong-Jian Han",
        "Chuan-Feng Li",
        "Guang-Can Guo"
      ],
      "abstract": "Pure-state manifestations of geometric phase are well established and have found applications across essentially all branches of physics, yet their generalization to mixed-state regimes remains largely unexplored experimentally. The Uhlmann geometric phase offers a natural extension of pure-state paradigms and can exhibit a topological character. However, observation of this invariant is impeded by the incompatibility between Uhlmann parallel transport and Hamiltonian dynamics, as well as the difficulty of preparing topologically nontrivial mixed states. To address this challenge, we report an experimentally accessible protocol for directly measuring the mixed-state topological invariant. By engineering controlled nonunitary dynamics in a photonic quantum walk, we prepare topologically nontrivial mixed states from a trivial initial state. Furthermore, by machine-learning the full density matrix in momentum space, we directly extract the quantized geometric phase of the nontrivial mixed states. These results highlight a geometric phase framework that naturally extends to open quantum systems both in and out of thermal equilibrium.",
      "url": "http://arxiv.org/abs/2512.24857v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24857v1",
      "citations": null,
      "categories": [
        "quant-ph",
        "physics.optics"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 537
    },
    {
      "title": "Advances in Agentic AI: Back to the Future",
      "authors": [
        "Sergio Alvarez-Telena",
        "Marta Diez-Fernandez"
      ],
      "abstract": "In light of the recent convergence between Agentic AI and our field of Algorithmization, this paper seeks to restore conceptual clarity and provide a structured analytical framework for an increasingly fragmented discourse. First, (a) it examines the contemporary landscape and proposes precise definitions for the key notions involved, ranging from intelligence to Agentic AI. Second, (b) it reviews our prior body of work to contextualize the evolution of methodologies and technological advances developed over the past decade, highlighting their interdependencies and cumulative trajectory. Third, (c) by distinguishing Machine and Learning efforts within the field of Machine Learning (d) it introduces the first Machine in Machine Learning (M1) as the underlying platform enabling today's LLM-based Agentic AI, conceptualized as an extension of B2C information-retrieval user experiences now being repurposed for B2B transformation. Building on this distinction, (e) the white paper develops the notion of the second Machine in Machine Learning (M2) as the architectural prerequisite for holistic, production-grade B2B transformation, characterizing it as Strategies-based Agentic AI and grounding its definition in the structural barriers-to-entry that such systems must overcome to be operationally viable. Further, (f) it offers conceptual and technical insight into what appears to be the first fully realized implementation of an M2. Finally, drawing on the demonstrated accuracy of the two previous decades of professional and academic experience in developing the foundational architectures of Algorithmization, (g) it outlines a forward-looking research and transformation agenda for the coming two decades.",
      "url": "http://arxiv.org/abs/2512.24856v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24856v1",
      "citations": null,
      "categories": [
        "econ.TH",
        "cs.AR",
        "cs.CE",
        "cs.ET"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 538
    },
    {
      "title": "QCD Wehrl and entanglement entropies in a gluon spectator model at small-$x$",
      "authors": [
        "Gabriel Rabelo-Soares",
        "Reinaldo Francener",
        "Gabriel S. Ramos",
        "Giorgio Torrieri"
      ],
      "abstract": "Recent studies have shown that hadronic multiplicity in deep inelastic scattering is associated with an entanglement entropy. However, such definitions are intrinsically longitudinal and do not capture the full phase--space structure of the proton. In this work, we investigate the Wehrl entropy of the proton constructed from Husimi distribution obtained from the Gaussian smearing of the Wigner distribution. We show that the entanglement entropy naturally emerges from the normalization condition of the Husimi distribution within this framework. In addition, the Wehrl entropy contains a contribution associated with transverse degrees of freedom. Numerical results for the proton Wehrl entropy are presented for different values of the virtuality.",
      "url": "http://arxiv.org/abs/2512.24855v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24855v1",
      "citations": null,
      "categories": [
        "hep-ph",
        "nucl-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 539
    },
    {
      "title": "Manifold-Constrained Sentence Embeddings via Triplet Loss: Projecting Semantics onto Spheres, Tori, and Möbius Strips",
      "authors": [
        "Vinit K. Chavan"
      ],
      "abstract": "Recent advances in representation learning have emphasized the role of embedding geometry in capturing semantic structure. Traditional sentence embeddings typically reside in unconstrained Euclidean spaces, which may limit their ability to reflect complex relationships in language. In this work, we introduce a novel framework that constrains sentence embeddings to lie on continuous manifolds -- specifically the unit sphere, torus, and M\\\"obius strip -- using triplet loss as the core training objective. By enforcing differential geometric constraints on the output space, our approach encourages the learning of embeddings that are both discriminative and topologically structured. We evaluate our method on benchmark datasets (AG News and MBTI) and compare it to classical baselines including TF-IDF, Word2Vec, and unconstrained Keras-derived embeddings. Our results demonstrate that manifold-constrained embeddings, particularly those projected onto spheres and M\\\"obius strips, significantly outperform traditional approaches in both clustering quality (Silhouette Score) and classification performance (Accuracy). These findings highlight the value of embedding in manifold space -- where topological structure complements semantic separation -- offering a new and mathematically grounded direction for geometric representation learning in NLP.",
      "url": "https://www.semanticscholar.org/paper/04bc6f7d6e660b62e7f55a309e24e7bd87e9eb88",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2505.00014",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 540
    },
    {
      "title": "Dynamical Geometric Theory of Principal Bundle Constrained Systems: Strong Transversality Conditions and Variational Framework for Gauge Field Coupling",
      "authors": [
        "Dongzhe Zheng"
      ],
      "abstract": "This paper introduces a geometric mechanics framework for constrained systems on principal bundles through \\emph{compatible pairs} $(\\mathcal{D}, \\lambda)$, addressing fundamental challenges in gauge-constrained physical systems. We characterize the strong transversality condition by pairing constraint distributions $\\mathcal{D}$ with Lie algebra dual functions $\\lambda: P \\to \\mathfrak{g}^*$ satisfying compatibility $\\mathcal{D}_p = \\{v : \\langle\\lambda(p), \\omega(v)\\rangle = 0\\}$ and differential consistency $d\\lambda + \\mathrm{ad}^*_\\omega \\lambda = 0$. This framework proves equivalent to $G$-equivariant Atiyah sequence splittings. We establish bidirectional construction enabling computation: forward (from $\\lambda$ to compatible $\\mathcal{D}$) and inverse (via variational minimization). Key mathematical contributions include existence theorems for bundles with $\\mathrm{ad}^*_\\Omega\\lambda = 0$ and uniqueness results for semi-simple groups with $\\mathfrak{z}(\\mathfrak{g}) = 0$, providing rigorous foundations for constraint classification. Our central physical insight emerges from variational principles, deriving dynamic connection equations $\\partial_t\\omega = d^{\\omega}\\eta - \\iota_{X_H}\\Omega$ revealing constraint-curvature coupling: $P_{\\text{constraint}} = \\langle\\lambda,\\Omega(\\dot{q},\\delta q)\\rangle$. This explains non-trivial constraint-field interactions in magnetohydrodynamics and Yang-Mills systems absent in kinematic theories. We construct Spencer cohomology for compatible pairs, establishing deep connections between topological invariants and conservation laws. Systematic comparison demonstrates that strong transversality captures essential constraint-curvature physics invisible to standard approaches. Applications span fluid dynamics, gauge theories, and geometric control, providing new tools for complex physical systems with intrinsic gauge-constraint coupling.",
      "url": "https://www.semanticscholar.org/paper/07b3d3c138b8c41624a4cfcadfb103266dc8040b",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 6,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 541
    },
    {
      "title": "Computation of connection-based Zagreb indices in chain graphs and triangular sheets",
      "authors": [
        "Muhammad Mudassar Hassan",
        "A. Waqar",
        "Haidar Ali",
        "Parvez Ali"
      ],
      "abstract": "Abstract Graph theory is a mathematical framework that can be used to model and analyze complex networks. Topology plays a key role in determining the compatibility of chemical graphs. Topological indices have been widely applied in a wide range of domains, such as chemistry, biological activity prediction, environmental risk assessment, and drug design. These indices can predict aqueous solubility, dipole moment, viscosity, boiling degree, refractive index, toxicity, and surface tension, among other physicochemical characteristics. Topological descriptors also assist us in the quantitative analysis and comparison of molecules by simplifying complicated molecular information. In this paper, we have computed the first Zagreb connection index ZC1, the second Zagreb connection index ZC2, and the first modified Zagreb connection index ZC1*. Further, a few recently introduced Zagreb connection indices like atom-bond connectivity, geomatric-arithmatic, and hyper-Zagreb connection indices are also calculated for the melem chain MC(m), the borophene chain B36(m), and the boron triangular sheet BTS(m, n). Our goal is to comprehend the molecular connections between these substances. Through the study of several mathematical methods, we discovered intriguing relationships and patterns within their structures. The primary objective of our research is to learn more about these compounds and the relationship between their molecular structures and properties. Every chemical, as we have seen, has distinct properties. The ability to predict the behavior of these molecules under different conditions might open up new research avenues and practical applications.",
      "url": "https://www.semanticscholar.org/paper/c4678c934de620dc5321105ec4a88f8207dd6604",
      "year": 2024,
      "venue": "Journal of coordination chemistry",
      "source": "semantic_scholar",
      "doi": "10.1080/00958972.2024.2305819",
      "pdf_url": "",
      "citations": 5,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 542
    },
    {
      "title": "Numerical Simulations and Bifurcation of Ca2+ Oscillatory Behaviour in the Connection of Neurons and Astrocytes",
      "authors": [
        "Hemlata Jethanandani̇",
        "B. Jha"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/9ff019beb4e7a5e1d7d871088d3ad0d18265eacb",
      "year": 2024,
      "venue": "Cell Biochemistry and Biophysics",
      "source": "semantic_scholar",
      "doi": "10.1007/s12013-024-01427-1",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 543
    },
    {
      "title": "Constrained Branching Search for Topology Identification Stream Computing With Lightweight Implementation",
      "authors": [
        "Zhuoheng Wang",
        "Jie Gao",
        "Qiushi Cui",
        "Yang Weng"
      ],
      "abstract": "Accurate topological awareness is critical to the stability of low-voltage distribution networks (LVDNs). However, traditional impedance-based topology restoration assumes accuracy that is often unattainable due to impedance data inaccuracy. Given LVDN sensor quality, robustness against data quality issue is crucial. Additionally, the integration of distributed energy resources (DERs) is expanding. Identifying their locations is necessary for effective load management and decreasing utility loss. Conventional identification methods rely on centralized data processing. However, they are limited due to increased storage and computational demands. This paper presents a novel approach employing constrained branching search within a stream computing framework, tailored for radial LVDNs. The proposed method uses node connection (NC) restrictions to recover topology. These constraints are based on the radial LVDN physical model. Additionally, a mathematical model for plug-in PV locations is introduced. We design CommuniDispatch, a lightweight implementation stream computing framework integrating our topology identification method. Enhanced by a Latin hypercube sampling-based recursive bound & search (LHS-RBS) algorithm, it significantly amplifies computational efficiency. Our experiments on diverse radial LVDNs validate the method's accuracy in topology identification and robustness against data quality issues, along with plug-in PV location and the computational efficiency of the LHS-RBS.",
      "url": "https://www.semanticscholar.org/paper/91b6244b4f5ed43ea339bacdf1b4eac2ac3acf87",
      "year": 2025,
      "venue": "IEEE Transactions on Power Systems",
      "source": "semantic_scholar",
      "doi": "10.1109/TPWRS.2024.3510940",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 544
    },
    {
      "title": "A Differential Manifold Perspective and Universality Analysis of Continuous Attractors in Artificial Neural Networks",
      "authors": [
        "Shaoxin Tian",
        "Hongkai Liu",
        "Yuying Yang",
        "Jiali Yu",
        "Zizheng Miao",
        "Xuming Huang",
        "Zhishuai Liu",
        "Zhang Yi"
      ],
      "abstract": "Continuous attractors are critical for information processing in both biological and artificial neural systems, with implications for spatial navigation, memory, and deep learning optimization. However, existing research lacks a unified framework to analyze their properties across diverse dynamical systems, limiting cross-architectural generalizability. This study establishes a novel framework from the perspective of differential manifolds to investigate continuous attractors in artificial neural networks. It verifies compatibility with prior conclusions, elucidates links between continuous attractor phenomena and eigenvalues of the local Jacobian matrix, and demonstrates the universality of singular value stratification in common classification models and datasets. These findings suggest continuous attractors may be ubiquitous in general neural networks, highlighting the need for a general theory, with the proposed framework offering a promising foundation given the close mathematical connection between eigenvalues and singular values.",
      "url": "https://www.semanticscholar.org/paper/442477ffcc9360823bd8a0b47c9b13c30ad19da1",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2509.10514",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 545
    },
    {
      "title": "Hyper-differential sensitivity analysis for inverse problems constrained by partial differential equations",
      "authors": [
        "Isaac Sunseri",
        "Joseph L. Hart",
        "Bart van Bloemen Waanders",
        "A. Alexanderian"
      ],
      "abstract": "High fidelity models used in many science and engineering applications couple multiple physical states and parameters. Inverse problems arise when a model parameter cannot be determined directly, but rather is estimated using (typically sparse and noisy) measurements of the states. The data is usually not sufficient to simultaneously inform all of the parameters. Consequently, the governing model typically contains parameters which are uncertain but must be specified for a complete model characterization necessary to invert for the parameters of interest. We refer to the combination of the additional model parameters (those which are not inverted for) and the measured data states as the ‘complementary parameters’. We seek to quantify the relative importance of these complementary parameters to the solution of the inverse problem. To address this, we present a framework based on hyper-differential sensitivity analysis (HDSA). HDSA computes the derivative of the solution of an inverse problem with respect to complementary parameters. We present a mathematical framework for HDSA in large-scale PDE-constrained inverse problems and show how HDSA can be interpreted to give insight about the inverse problem. We demonstrate the effectiveness of the method on an inverse problem by estimating a permeability field, using pressure and concentration measurements, in a porous medium flow application with uncertainty in the boundary conditions, source injection, and diffusion coefficient.",
      "url": "https://www.semanticscholar.org/paper/6dee5be22510e13a7d228b3ebb9cb85f650df7f7",
      "year": 2020,
      "venue": "Inverse Problems",
      "source": "semantic_scholar",
      "doi": "10.1088/1361-6420/abaf63",
      "pdf_url": "https://arxiv.org/pdf/2003.00978",
      "citations": 23,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 546
    },
    {
      "title": "Advancing Mathematical Research via Human-AI Interactive Theorem Proving",
      "authors": [
        "Chenyi Li",
        "Zhijian Lai",
        "Dong An",
        "Jiang Hu",
        "Zaiwen Wen"
      ],
      "abstract": "We investigate how large language models can be used as research tools in scientific computing while preserving mathematical rigor. We propose a human-in-the-loop workflow for interactive theorem proving and discovery with LLMs. Human experts retain control over problem formulation and admissible assumptions, while the model searches for proofs or contradictions, proposes candidate properties and theorems, and helps construct structures and parameters that satisfy explicit constraints, supported by numerical experiments and simple verification checks. Experts treat these outputs as raw material, further refine them, and organize the results into precise statements and rigorous proofs. We instantiate this workflow in a case study on the connection between manifold optimization and Grover's quantum search algorithm, where the pipeline helps identify invariant subspaces, explore Grover-compatible retractions, and obtain convergence guarantees for the retraction-based gradient method. The framework provides a practical template for integrating large language models into frontier mathematical research, enabling faster exploration of proof space and algorithm design while maintaining transparent reasoning responsibilities. Although illustrated on manifold optimization problems in quantum computing, the principles extend to other core areas of scientific computing.",
      "url": "https://www.semanticscholar.org/paper/3fdfca6bfbb95d382ac285fdef221592c9f8ee83",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 547
    },
    {
      "title": "Global thermodynamic manifold for conservative control of stochastic systems",
      "authors": [
        "Jordan R. Sawchuk",
        "David A. Sivak"
      ],
      "abstract": "Optimal control of stochastic systems plays a central role in nonequilibrium physics, with applications in the study of biological molecular motors and the design of single-molecule experiments. While exact analytic solutions to optimization problems are rare, under slow driving conditions, the problem can be reformulated geometrically solely in terms of equilibrium properties. In this framework, minimum-work protocols are geodesics on a thermodynamic manifold whose metric is a generalized friction tensor. Here, we introduce a new foundation for this friction-tensor formalism for conservatively driven systems. Under complete control of the potential energy, a global thermodynamic manifold (on which points are identified with instantaneous energy landscapes) has as its metric a full-control friction tensor. Arbitrary partial-control friction tensors arise naturally as inherited metrics on submanifolds of this global manifold. Leveraging a simple mathematical relationship between system dynamics and the geometry of the global manifold, we derive new expressions for the friction tensor that offer powerful tools for interpretation and computation of friction tensors and minimum-work protocols. Our results elucidate a connection between relaxation and dissipation in slowly driven systems and suggest optimization heuristics. We demonstrate the utility of these developments in three illustrative examples.",
      "url": "https://www.semanticscholar.org/paper/430a0e42e4a0350783cdac1c76b7ec4682914610",
      "year": 2024,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 548
    },
    {
      "title": "DisCO: Reinforcing Large Reasoning Models with Discriminative Constrained Optimization",
      "authors": [
        "Gang Li",
        "Ming Lin",
        "Tomer Galanti",
        "Zhengzhong Tu",
        "Tianbao Yang"
      ],
      "abstract": "The recent success and openness of DeepSeek-R1 have brought widespread attention to Group Relative Policy Optimization (GRPO) as a reinforcement learning method for large reasoning models (LRMs). In this work, we analyze the GRPO objective under a binary reward setting and reveal an inherent limitation of question-level difficulty bias. We also identify a connection between GRPO and traditional discriminative methods in supervised learning. Motivated by these insights, we introduce a new Discriminative Constrained Optimization (DisCO) framework for reinforcing LRMs, grounded in the principle of discriminative learning. The main differences between DisCO and GRPO and its recent variants are: (1) it replaces the group relative objective with a discriminative objective defined by a scoring function; (2) it abandons clipping-based surrogates in favor of non-clipping RL surrogate objectives used as scoring functions; (3) it employs a simple yet effective constrained optimization approach to enforce the KL divergence constraint. As a result, DisCO offers notable advantages over GRPO and its variants: (i) it completely eliminates difficulty bias by adopting discriminative objectives; (ii) it addresses the entropy instability in GRPO and its variants through the use of non-clipping scoring functions and a constrained optimization approach, yielding long and stable training dynamics; (iii) it allows the incorporation of advanced discriminative learning techniques to address data imbalance, where a significant number of questions have more negative than positive generated answers during training. Our experiments on enhancing the mathematical reasoning capabilities of SFT-finetuned models show that DisCO significantly outperforms GRPO and its improved variants such as DAPO, achieving average gains of 7\\% over GRPO and 6\\% over DAPO across six benchmark tasks for an 1.5B model.",
      "url": "https://www.semanticscholar.org/paper/2ab138c8ad6be7bfa4c001fdc51232e687c878be",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2505.12366",
      "pdf_url": "",
      "citations": 7,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 549
    },
    {
      "title": "Transformers from Diffusion: A Unified Framework for Neural Message Passing",
      "authors": [
        "Qitian Wu",
        "David Wipf",
        "Junchi Yan"
      ],
      "abstract": "Learning representations for structured data with certain geometries (e.g., observed or unobserved) is a fundamental challenge, wherein message passing neural networks (MPNNs) have become a de facto class of model solutions. In this paper, inspired by physical systems, we propose an energy-constrained diffusion model, which integrates the inductive bias of diffusion on manifolds with layer-wise constraints of energy minimization. We identify that the diffusion operators have a one-to-one correspondence with the energy functions implicitly descended by the diffusion process, and the finite-difference iteration for solving the energy-constrained diffusion system induces the propagation layers of various types of MPNNs operating on observed or latent structures. This leads to a unified mathematical framework for common neural architectures whose computational flows can be cast as message passing (or its special case), including MLPs, GNNs, and Transformers. Building on these insights, we devise a new class of neural message passing models, dubbed diffusion-inspired Transformers (DIFFormer), whose global attention layers are derived from the principled energy-constrained diffusion framework. Across diverse datasets ranging from real-world networks to images, texts, and physical particles, we demonstrate that the new model achieves promising performance in scenarios where the data structures are observed (as a graph), partially observed, or entirely unobserved.",
      "url": "https://www.semanticscholar.org/paper/d545bf74d07796e614eb41c540a5bf4a8d151e62",
      "year": 2024,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 550
    },
    {
      "title": "Digital Construction Planning for Resource-Constrained Projects",
      "authors": [
        "Yan Lin"
      ],
      "abstract": "Aiming at the problem of construction period delay caused by unreasonable initial construction period design and schedule arrangement in the construction process of engineering projects, this paper explores a digital construction planning method for resource-constrained projects based on the trade-off between construction period and resources. By combining digital information technology with construction technology, an automatic optimization model covering BIM technology, mathematical algorithms and engineering practice is constructed, which is intended to provide accurate decision support for construction planning and schedule arrangement. On one hand, a process logic library considering flow overlapping is built through knowledge or rule reasoning, and scattered construction experience is transformed into structured logical rules to provide underlying framework logic support for digital construction planning. On the other hand, based on the conventional quota method, the trade-off analysis between construction period and resources and iterative optimization calculation are carried out. Taking meeting the contract construction period requirements as the basis, by reasonably dividing the flow sections and setting an appropriate flow rhythm, the minimization of resource demand is achieved and the efficiency of construction organization is improved under the premise of ensuring continuous on-site operation and orderly connection.",
      "url": "https://www.semanticscholar.org/paper/fed40af67db73ee4a21c66a79b182b9695494b9a",
      "year": 2025,
      "venue": "Proceedings of the 2025 8th International Conference on Computer Information Science and Artificial Intelligence",
      "source": "semantic_scholar",
      "doi": "10.1145/3773365.3773545",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 551
    },
    {
      "title": "Graph-MARL: A Neuro-Symbolic Autonomy Framework for Multi-Chaser Active Debris Removal Task Allocation and Path Planning",
      "authors": [
        "Dandan Su",
        "K. A. Neusypin",
        "Ge Dong"
      ],
      "abstract": "The escalating density of objects in Low Earth Orbit (LEO) renders Active Debris Removal (ADR) critical for ensuring the safety of space assets and the sustainability of the orbital environment. However, planning missions for a multi-chaser spacecraft constellation presents a complex, Astro dynamically-constrained combinatorial optimization problem, where conventional methods face significant challenges in scalability, non-stationarity, and handling non-Euclidean cost functions. This paper presents a rigorous mathematical framework for the Graph-enhanced Multi-Agent Reinforcement Learning (Graph-MARL) system designed for Active Debris Removal (ADR) missions. We systematically derive the mathematical foundations underlying each component, elucidating the interconnections between Graph Attention Networks (GAT) for relational reasoning, Multi-Agent Proximal Policy Optimization (MAPPO) for decentralized decision-making, and orbital dynamics constraints. The framework bridges the gap between abstract mathematical formulations and practical implementation requirements, providing explicit connections between theoretical constructs and their computational realizations.",
      "url": "https://www.semanticscholar.org/paper/8b82ea957d7ce5afca46743618ebfda5d5dc0567",
      "year": 2025,
      "venue": "2025 IEEE International Conference on Unmanned Systems (ICUS)",
      "source": "semantic_scholar",
      "doi": "10.1109/ICUS66297.2025.11294460",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 552
    },
    {
      "title": "The Connection on Fiber Bundles",
      "authors": [
        "Abd Rahman"
      ],
      "abstract": "A connection is a device that defines the concept of parallel transport on a bundle, that is identifies fibers over nearby points. Fiber bundles form are the natural mathematical framework for the gauge filed theories. Also affine connection is the most elementary type of connection, a means of parallel transfer of tangent vectors on a manifold from one point to another. In any manifold with a positive dimension there is an infinite number of the affine connection; junctions are among the simplest methods to determine the differentiation of sections of vector bundles. Our goal in this paper is to identify the concept of connection in fiber bundles. We followed the analytical historical mathematical method and we found that the connection on the fiber bundle is a smooth distribution over the total bundle area, which is of central importance in modern geometry and leads to appropriate formulas for geometry constants.",
      "url": "https://www.semanticscholar.org/paper/824c048a2cabf14d9959d608f3b23893148b97d7",
      "year": 2022,
      "venue": "INTERNATIONAL JOURNAL OF MATHEMATICS AND COMPUTER RESEARCH",
      "source": "semantic_scholar",
      "doi": "10.47191/ijmcr/v10i7.04",
      "pdf_url": "https://ijmcr.in/index.php/ijmcr/article/download/438/367",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 553
    },
    {
      "title": "Modeling Molecular Interactions with Hyper-Networks and Super-Hyper-networks",
      "authors": [
        "Takaaki Fujita",
        "Muhammad Gulistan",
        "Arkan A. Ghaib"
      ],
      "abstract": "Graph theory examines the structure of networks by treating entities as vertices and the connections between them as edges. A Hyper-Graph enhances this framework by permitting a single Hyper-edge to link multiple vertices at once. Building on that idea, a Super-Hyper-Graph introduces layers of recursively nested powersets, which create hierarchical and self-referential relationships among its Hyper-edges. These extended models—often called Hyper-Networks and Super-Hyper-networks—capture complex, higher-order associations that ordinary graphs cannot. Such constructions find important applications in the life sciences. For example, a Molecular Interaction Network represents biochemical systems by assigning each molecule to a node and using edges to denote pairwise interactions or chemical reactions, thereby facilitating the analysis of intricate molecular pathways.\nIn this paper, we extend the concept of Molecular Interaction Networks by proposing two new frameworks: the Molecular Interaction Hyper-network and the Molecular Interaction Super-Hyper-network, both grounded in the structures of Hyper-Networks and Super-Hyper-networks. These frameworks offer new insights into multi-scale biochemical systems, with potential applications in drug target identification and pathway analysis. We hope that future research will further explore the mathematical, biological, and computational aspects of the Molecular Interaction Hyper-network and the Molecular Interaction Super-Hyper-network.",
      "url": "https://www.semanticscholar.org/paper/7658e365a780ca89498c140a52ca4af3ee90dcac",
      "year": 2025,
      "venue": "Advances in Research",
      "source": "semantic_scholar",
      "doi": "10.9734/air/2025/v26i41412",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 554
    },
    {
      "title": "Dynamic AP-UE Association and Power Allocation in Sparse LSFD for Energy-Constrained Networks",
      "authors": [
        "Shaik Karimullah",
        "S. Fahimuddin",
        "T. N. Ranganadham",
        "P. Hariobulesu",
        "R.Soma Sekhar",
        "C.Raviteja Reddy",
        "P. Shreya",
        "P. Revathi"
      ],
      "abstract": "The rise of 5G and beyond has intensified the need for energy-efficient, high-capacity wireless networks, especially in IoT, smart cities, and dense urban settings. Sparse Large-Scale Fading Decoding (S-LSFD) addresses this challenge by dynamically associating User Equipment (UE) with a selective subset of Access Points (APs) offering the strongest channel gains. This selective association reduces interference, improves spectral efficiency, and minimizes energy consumption while maintaining Quality of Service (QoS).S-LSFD incorporates optimized power allocation to focus energy on high-quality AP connections, maximizing transmission efficiency. Its adaptability is governed by the control parameter λ, which balances reliability and energy savings by adjusting the number of APs each UE connects to. This flexibility makes S-LSFD ideal for energy-constrained scenarios like IoT ecosystems and densely populated areas.This study explores dynamic AP-UE association and power allocation mechanisms in S-LSFD using mathematical modeling and real-world scenarios. The proposed framework ensures scalability, energy efficiency, and high data rates, providing a sustainable solution for next-generation wireless networks.",
      "url": "https://www.semanticscholar.org/paper/41096550bcaba7f47e0521077d153f955112600f",
      "year": 2025,
      "venue": "2025 International Conference on Computer, Electrical & Communication Engineering (ICCECE)",
      "source": "semantic_scholar",
      "doi": "10.1109/ICCECE61355.2025.10941057",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 555
    },
    {
      "title": "Physics-constrained normalizing flow for identification and modeling of vortex-induced vibration in stay cables",
      "authors": [
        "Zhe Wang",
        "Zhiping Mao",
        "Shanwu Li",
        "Yongchao Yang"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/6deff5669b35c03cc76d9aca92d728d767457d59",
      "year": 2025,
      "venue": "Nonlinear dynamics",
      "source": "semantic_scholar",
      "doi": "10.1007/s11071-025-11808-7",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 556
    },
    {
      "title": "Stochastic Mechanics: The Unification of Quantum Mechanics with Brownian Motion",
      "authors": [
        "F. Kuipers"
      ],
      "abstract": "We unify Brownian motion and quantum mechanics in a single mathematical framework. In particular, we show that non-relativistic quantum mechanics of a single spinless particle on a flat space can be described by a Wiener process that is rotated in the complex plane. We then extend this theory to relativistic stochastic theories on manifolds using the framework of second order geometry. As a byproduct, our results suggest that a consistent path integral based formulation of a quantum theory on a Lorentzian (Riemannian) manifold requires an Ito deformation of the Poincare (Galilean) symmetry, arising due to the coupling of the quadratic variation to the affine connection.",
      "url": "https://www.semanticscholar.org/paper/2c992e3b47c4b77f633cb50ad938dd6a4612f75b",
      "year": 2023,
      "venue": "SpringerBriefs in Physics",
      "source": "semantic_scholar",
      "doi": "10.1007/978-3-031-31448-3",
      "pdf_url": "",
      "citations": 19,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 557
    },
    {
      "title": "Geo-PhysNet: A Geometry-Aware and Physics-Constrained Graph Neural Network for Aerodynamic Pressure Prediction on Vehicle Fluid–Solid Surfaces",
      "authors": [
        "Bowen Liu",
        "Hao Wang",
        "Liheng Xue",
        "Yin Long"
      ],
      "abstract": "The aerodynamic pressure of a car is crucial for its shape design. To overcome the time-consuming and costly bottleneck of wind tunnel tests and computational fluid dynamics (CFD) simulations, deep learning-based surrogate models have emerged as highly promising alternatives. However, existing methods that only predict on the surface of objects only learn the mapping of pressure. In contrast, a physically realistic field has values and gradients that are structurally unified and self-consistent. Therefore, existing methods ignore the crucial differential structure and intrinsic continuity of the physical field as a whole. This oversight leads to their predictions, even if locally numerically close, often showing unrealistic gradient distributions and high-frequency oscillations macroscopically, greatly limiting their reliability and practicality in engineering decisions. To address this, this study proposes the Geo-PhysNet model, a graph neural network framework specifically designed for complex surface manifolds with strong physical constraints. This framework learns a differential representation, and its network architecture is designed to simultaneously predict the pressure scalar field and its tangential gradient vector field on the surface manifold within a unified framework. By making the gradient an explicit learning target, we force the network to understand the local mechanical causes leading to pressure changes, thereby mathematically ensuring the self-consistency of the field’s intrinsic structure, rather than merely learning the numerical mapping of pressure. Finally, to solve the common noise problem in the predictions of existing methods, we introduce a physical regularization term based on the surface Laplacian operator to penalize non-smooth solutions, ensuring the physical rationality of the final output field. Experimental verification results show that Geo-PhysNet not only outperforms existing benchmark models in numerical accuracy but, more importantly, demonstrates superior advantages in the physical authenticity, field continuity, and gradient smoothness of the generated pressure fields.",
      "url": "https://www.semanticscholar.org/paper/39d2ee645bea2279e81f9c393aa9f92513578f97",
      "year": 2025,
      "venue": "Applied Sciences",
      "source": "semantic_scholar",
      "doi": "10.3390/app152111645",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 558
    },
    {
      "title": "Local Thermal Operations and Classical Communication",
      "authors": [
        "Rafal Bistro'n",
        "Jakub Czartowski"
      ],
      "abstract": "In quantum thermodynamics, understanding the interplay between locality, thermal constraints, and communication remains an open challenge. In this manuscript, we introduce Local Thermal Operations and Classical Communication (LTOCC), a novel operational framework that unifies the distant laboratories paradigm with thermodynamic restrictions, defining the fundamental limits on transformations between spatially separated systems. We establish a hierarchy of LTOCC protocols, demonstrating inclusion relations between different levels and revealing their deep connection to semilocal thermal operations. To formalize this framework, we develop thermal tensors and bithermal tensors, extending stochastic and tristochastic tensors to thermodynamic settings and providing new mathematical tools for constrained quantum processes. Finally, we present limitations imposed by LTOCC on single- and multi-copy CHSH scenario, demonstrating no violation in former and a gap between thermal and athermal local operations in the latter with respect to their capability to detect entanglement.",
      "url": "https://www.semanticscholar.org/paper/9e296f97e03339c04b9ef8817f8c85d9866a09ae",
      "year": 2024,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 559
    },
    {
      "title": "A Formal Approach to Optimally Configure a Fully Connected Multilayer Hybrid Neural Network",
      "authors": [
        "Goutam Chakraborty",
        "V. Azhmyakov",
        "Luz Adriana Guzman Trujillo"
      ],
      "abstract": "This paper is devoted to a novel formal analysis, optimizing the learning models for feedforward multilayer neural networks with hybrid structures. The proposed mathematical description replicates a specific switched-type optimal control problem (OCP). We have developed an equivalent, optimal control-based formulation of the given problem of training a hybrid feedforward multilayer neural network, to train the target mapping function constrained by the training samples. This novel formal approach makes it possible to apply some well-established optimal control techniques to design a versatile type of full connection neural networks. We next discuss the irrelevance of the necessity of Pontryagin-type optimality conditions for the construction of the obtained switched-type OCP. This fact motivated us to consider the so-called direct-solution approaches to the switched OCPs, which can be associated with the learning of hybrid neural networks. Concretely, we consider the generalized reduced-gradient algorithm in the framework of the auxiliary switched OCP.",
      "url": "https://www.semanticscholar.org/paper/b3db2c99afe5331efd518378e6b8b52bdfca63d9",
      "year": 2024,
      "venue": "Mathematics",
      "source": "semantic_scholar",
      "doi": "10.3390/math13010129",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 560
    },
    {
      "title": "Nonlinear Dynamical Model and Analysis of Emotional Propagation Based on Caputo Derivative",
      "authors": [
        "Liang Hong",
        "Lipu Zhang"
      ],
      "abstract": "Conventional integer-order models fail to adequately capture non-local memory effects and constrained nonlinear interactions in emotional dynamics. To address these limitations, we propose a coupled framework that integrates Caputo fractional derivatives with hyperbolic tangent–based interaction functions. The fractional-order term quantifies power-law memory decay in affective states, while the nonlinear component regulates connection strength through emotional difference thresholds. Mathematical analysis establishes the existence and uniqueness of solutions with continuous dependence on initial conditions and proves the local asymptotic stability of network equilibria (Wij*=1δsech2(∥Ei−Ej∥), e.g., W*≈1.40 under typical parameters η=0.5, δ=0.3). We further derive closed-form expressions for the steady-state variance under stochastic perturbations (Var(Wij)=σζ22ηδ) and demonstrate a less than 6% deviation between simulated and theoretical values when σζ=0.1. Numerical experiments using the Euler–Maruyama method validate the convergence of connection weights toward the predicted equilibrium, reveal Gaussian features in the stationary distributions, and confirm power-law scaling between noise intensity and variance. The numerical accuracy of the fractional system is further verified through L1 discretization, with observed error convergence consistent with theoretical expectations for μ=0.5. This framework advances the mechanistic understanding of co-evolutionary dynamics in emotion-modulated social networks, supporting applications in clinical intervention design, collective sentiment modeling, and psychophysiological coupling research.",
      "url": "https://www.semanticscholar.org/paper/d36d8538a30af37158f785eeb34e739f31b62f96",
      "year": 2025,
      "venue": "Mathematics",
      "source": "semantic_scholar",
      "doi": "10.3390/math13132044",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 561
    },
    {
      "title": "Constrained Dynamics, Stochastic Numerical Methods and the Modeling of Complex Systems",
      "authors": [
        "B. Leimkuhler",
        "Richard Tsai",
        "Gilles Vilmart",
        "Rachel Ward"
      ],
      "abstract": "The workshop aimed to unite researchers from diverse fields of mathematics and statistics to explore the foundations of high-dimensional modeling and computational studies. It addressed recent advancements in numerical analysis, dynamical systems, and stochastic differential equations that support model reduction for large-scale complex systems.Incorporating targeted geometric structures, such as Riemannian manifolds, into large-scale statistical models is known to enhance the stability, reliability, and efficiency of numerical methods. However, algorithms are often presented in application contexts without adequate attention to their fundamental properties, limiting the adoption of these advanced modeling methods.The workshop emphasized understanding the fundamental properties of these structures, their impact on dynamics and stochastic dynamics, and the need to redesign algorithms to capture essential properties, aiming for robustness and suitability for high-performance computation.By bringing together numerical analysts, statisticians, and modelers, the workshop sought to improve the quality of methods and identify new model frameworks to guide future development.",
      "url": "https://www.semanticscholar.org/paper/84dad5d0ba35822145bb9df6cd454b3f4ef35726",
      "year": 2024,
      "venue": "Oberwolfach Reports",
      "source": "semantic_scholar",
      "doi": "10.4171/owr/2024/26",
      "pdf_url": "https://ems.press/content/serial-article-files/49484",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 562
    },
    {
      "title": "Hyper-differential sensitivity analysis with respect to model discrepancy: mathematics and computation",
      "authors": [
        "Joseph L. Hart",
        "B. V. B. Waanders"
      ],
      "abstract": "Model discrepancy, defined as the difference between model predictions and reality, is ubiquitous in computational models for physical systems. It is common to derive partial differential equations (PDEs) from first principles physics, but make simplifying assumptions to produce tractable expressions for the governing equations or closure models. These PDEs are then used for analysis and design to achieve desirable performance. For instance, the end goal may be to solve a PDE-constrained optimization (PDECO) problem. This article considers the sensitivity of PDECO problems with respect to model discrepancy. We introduce a general representation of the discrepancy and apply post-optimality sensitivity analysis to derive an expression for the sensitivity of the optimal solution with respect to the discrepancy. An efficient algorithm is presented which combines the PDE discretization, post-optimality sensitivity operator, adjoint-based derivatives, and a randomized generalized singular value decomposition to enable scalable computation. Kronecker product structure in the underlying linear algebra and corresponding infrastructure in PDECO is exploited to yield a general purpose algorithm which is computationally efficient and portable across a range of applications. Known physics and problem specific characteristics of discrepancy are imposed through user specified weighting matrices. We demonstrate our proposed framework on two nonlinear PDECO problems to highlight its computational efficiency and rich insight.",
      "url": "https://www.semanticscholar.org/paper/9b3975ee3ddcda50fca0321ed7889ca3f6fb97bf",
      "year": 2022,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2210.09037",
      "pdf_url": "https://arxiv.org/pdf/2210.09037",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 563
    },
    {
      "title": "On the Whitney near extension problem, BMO, alignment of data, best approximation in algebraic geometry, manifold learning and their beautiful connections: A modern treatment",
      "authors": [
        "S. Damelin"
      ],
      "abstract": "This paper provides fascinating connections between several mathematical problems which lie on the intersection of several mathematics subjects, namely algebraic geometry, approximation theory, complex-harmonic analysis and high dimensional data science. Modern techniques in algebraic geometry, approximation theory, computational harmonic analysis and extensions develop the first of its kind, a unified framework which allows for a simultaneous study of labeled and unlabeled near alignment data problems in of $\\mathbb R^D$ with the near isometry extension problem for discrete and non-discrete subsets of $\\mathbb R^D$ with certain geometries. In addition, the paper surveys related work on clustering, dimension reduction, manifold learning, vision as well as minimal energy partitions, discrepancy and min-max optimization. Numerous open problems are given.",
      "url": "https://www.semanticscholar.org/paper/00fa67298edfb51d69a830f539b3085351c545cf",
      "year": 2021,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 564
    },
    {
      "title": "Hyperspectral subpixel unmixing via an integrative framework",
      "authors": [
        "Chunzhi Li",
        "Xiaohua Chen",
        "Yuan Zhang"
      ],
      "abstract": "ABSTRACT In hyperspectral applications, spectral unmixing (SU) is an important technology to obtain the endmembers and the fractional land covers. Spectral variability, outliers, and nonlinearity are three challenging issues, causing SU to extract endmembers and corresponding abundance maps inaccurately. However, in view of the complexity of the three issues, to tackle them at once is difficult and intractable. In this paper, all the aforementioned issues are advocated to process together by a powerful integrative SU framework, where a hyper-manifold learning approach via a sparsity-constrained dual is exploited. In the proposed integrative SU framework, heterogeneous and homogeneous information is explored by dividing the hyperspectral data (HD) into a series of sub-blocks, hyper-Laplacian-based graph is employed to address the nonlinearity, and spectral variability is controlled by a scaled matrix. Moreover, the interferences of the outliers are handled by the augmented correntropy-induced metric (ACIM), where the rare endmembers are separated from the abrupt anomalies via decomposing the HD sets in a low-rank structure and constraining the sparsity on the corresponding residual term. The experimental results on several popularly used real HD sets indicate the superior performances of the proposed approach.",
      "url": "https://www.semanticscholar.org/paper/c947cdc476ae597671bb4d4f923118b60eadfefb",
      "year": 2020,
      "venue": "",
      "source": "semantic_scholar",
      "doi": "10.1080/01431161.2020.1783711",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 565
    },
    {
      "title": "From Barthel–Randers–Kropina Geometries to the Accelerating Universe: A Brief Review of Recent Advances in Finslerian Cosmology",
      "authors": [
        "A. Bouali",
        "H. Chaudhary",
        "L. Csillag",
        "Rattanasak Hama",
        "T. Harko",
        "S. Sabău",
        "S. Shahidi"
      ],
      "abstract": "We present a review of recent developments in cosmological models based on Finsler geometry, as well as geometric extensions of general relativity formulated within this framework. Finsler geometry generalizes Riemannian geometry by allowing the metric tensor to depend not only on position but also on an additional internal degree of freedom, typically represented by a vector field at each point of the spacetime manifold. We examine in detail the possibility that Finsler-type geometries can describe the physical properties of the gravitational interaction, as well as the cosmological dynamics. In particular, we present and review the implications of a particular implementation of Finsler geometry, based on the Barthel connection, and of the (α,β) geometries, where α is a Riemannian metric, and β is a one-form. For a specific construction of the deviation part β, in these classes of geometries, the Barthel connection coincides with the Levi–Civita connection of the associated Riemann metric. We review the properties of the gravitational field, and of the cosmological evolution in three types of geometries: the Barthel–Randers geometry, in which the Finsler metric function F is given by F=α+β, in the Barthel–Kropina geometry, with F=α2/β, and in the conformally transformed Barthel–Kropina geometry, respectively. After a brief presentation of the mathematical foundations of the Finslerian-type modified gravity theories, the generalized Friedmann equations in these geometries are written down by considering that the background Riemannian metric in the Randers and Kropina line elements is of Friedmann–Lemaitre–Robertson–Walker type. The matter energy balance equations are also presented, and they are interpreted from the point of view of the thermodynamics of irreversible processes in the presence of particle creation. We investigate the cosmological properties of the Barthel–Randers and Barthel–Kropina cosmological models in detail. In these scenarios, the additional geometric terms arising from the Finslerian structure can be interpreted as an effective geometric dark energy component, capable of generating an effective cosmological constant. Several cosmological solutions—both analytical and numerical—are obtained and compared against observational datasets, including Cosmic Chronometers, Type Ia Supernovae, and Baryon Acoustic Oscillations, using a Markov Chain Monte Carlo (MCMC) analysis. A direct comparison with the standard ΛCDM model is also carried out. The results indicate that Finslerian cosmological models provide a satisfactory fit to the observational data, suggesting they represent a viable alternative to the standard cosmological model based on general relativity.",
      "url": "https://www.semanticscholar.org/paper/eefe07bdf746f006400e52eccdc94d868a732831",
      "year": 2025,
      "venue": "Universe",
      "source": "semantic_scholar",
      "doi": "10.3390/universe11070198",
      "pdf_url": "",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 566
    },
    {
      "title": "Incompatible Deformations in Relativistic Elasticity",
      "authors": [
        "S. Lychev",
        "K. Koifman",
        "N. A. Pivovaroff"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/efb7876fd828ab8f3c15e2c39fa127cc5688b18b",
      "year": 2023,
      "venue": "Lobachevskii Journal of Mathematics",
      "source": "semantic_scholar",
      "doi": "10.1134/S1995080223060343",
      "pdf_url": "",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 567
    },
    {
      "title": "A Theory of Functional Connections-Based hp-Adaptive Mesh Refinement Algorithm for Solving Hypersensitive Two-Point Boundary-Value Problems",
      "authors": [
        "K. Drozd",
        "Roberto Furfaro",
        "Andrea D’Ambrosio"
      ],
      "abstract": "This manuscript introduces the first hp-adaptive mesh refinement algorithm for the Theory of Functional Connections (TFC) to solve hypersensitive two-point boundary-value problems (TPBVPs). The TFC is a mathematical framework that analytically satisfies linear constraints using an approximation method called a constrained expression. The constrained expression utilized in this work is composed of two parts. The first part consists of Chebyshev orthogonal polynomials, which conform to the solution of differentiation variables. The second part is a summation of products between switching and projection functionals, which satisfy the boundary constraints. The mesh refinement algorithm relies on the truncation error of the constrained expressions to determine the ideal number of basis functions within a segment’s polynomials. Whether to increase the number of basis functions in a segment or divide it is determined by the decay rate of the truncation error. The results show that the proposed algorithm is capable of solving hypersensitive TPBVPs more accurately than MATLAB R2021b’s bvp4c routine and is much better than the standard TFC method that uses global constrained expressions. The proposed algorithm’s main flaw is its long runtime due to the numerical approximation of the Jacobians.",
      "url": "https://www.semanticscholar.org/paper/cdd809a9e6328fced9561e0385680b88cb80a231",
      "year": 2024,
      "venue": "Mathematics",
      "source": "semantic_scholar",
      "doi": "10.3390/math12091360",
      "pdf_url": "https://www.mdpi.com/2227-7390/12/9/1360/pdf?version=1714406956",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 568
    },
    {
      "title": "Plato’s Allegory of the ‘Cave’ and Hyperspaces: Sonic Representation of the ‘Cave’ as a Four-Dimensional Acoustic Space via an Interactive Art Application",
      "authors": [
        "Dimitrios Traperas",
        "Andreas Floros",
        "N. Kanellopoulos"
      ],
      "abstract": "Mathematician and philosopher Charles Howard Hinton posited a plausible correlation between higher-dimensional spaces, also referred to as ‘hyperspaces’, and the allegorical concept articulated by the Ancient Greek philosopher Plato in his work, Republic, known as the ‘Cave.’ In Plato’s allegory, individuals find themselves situated in an underground ‘Cave’, constrained by chains on their legs and neck, perceiving shadows and sound reflections from the ‘real’ world cast on the ‘Cave’ wall as their immediate reality. Hinton extended the interpretation of these ‘shadows’ through the induction method, asserting that, akin to a 3D object casting a 2D shadow, the ‘shadow’ of a 4D hyper-object would exhibit one dimension less, manifesting as a 3D object. Building upon this conceptual framework, the authors posit a correlation between the perceived acoustic space of the bounded individuals within the ‘Cave’ and the characteristics of a 4D acoustic space, a proposition substantiated mathematically by scientific inquiry. Furthermore, the authors introduce an interactive art application developed as a methodical approach to exploring the hypothetical 4D acoustic space within Plato’s ‘Cave’, as perceived by the bounded individuals and someone liberated from his constraints navigating through the ‘Cave.’",
      "url": "https://www.semanticscholar.org/paper/58d9c906630243e25ed34ef3ce54f6796bbc4a4f",
      "year": 2024,
      "venue": "AppliedMath",
      "source": "semantic_scholar",
      "doi": "10.3390/appliedmath4030052",
      "pdf_url": "https://www.mdpi.com/2673-9909/4/3/52/pdf?version=1724299490",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 569
    },
    {
      "title": "Higgs Branches in the Omega-background via the Category of Line Operators",
      "authors": [
        "Thomas Karabela",
        "Wenjun Niu"
      ],
      "abstract": "The vacuum manifold $\\mathcal{M}$ of a topological twist of a 3d $\\mathcal{N}=4$ gauge theory is a hyper-K\\\"ahler variety; deformations and quantizations of $\\mathcal{M}$ can be constructed in the framework of 3 dimensional topological quantum field theories. In particular, based on physics arguments, turning on an omega-background results in the quantization of $\\mathbb{C}[\\mathcal{M}]$ as a Poisson algebra. In this paper, we implement this idea mathematically, in the context of the B-twist of 3d $\\mathcal{N}=4$ gauge theories, namely for Higgs branches. Our strategy is to implement omega-background in the category of line operators via the choice of a ribbon twist, and obtain the quantum Hamiltonian reduction as derived endomorphisms in the equivariant category, the category where the ribbon twist acts trivially. We apply quadratic Koszul duality to perform this computation.",
      "url": "https://www.semanticscholar.org/paper/e1cf61cde59a817e46664a18d544015485fdb82c",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 570
    },
    {
      "title": "Topology and knot theory applications in quantum field theory",
      "authors": [
        "Chaitali Deshpande",
        "Chandrashekhar Ramtirthkar",
        "T. A. Wani",
        "Varsha Kiran Bhosale",
        "Monali Gulhane",
        "K. S. Kumar"
      ],
      "abstract": "Topology and knot theory have profoundly influenced modern quantum field theory by revealing deep connections between gauge fields and topological invariants. Topological quantum field theories (TQFTs) disregard the spacetime metric and compute quantities that depend only on the global topology of the manifold. In Chern–Simons theory, observables associated with knotted loops correspond to knot polynomials such as the Jones polynomial. This paper reviews the mathematical framework that relates knots to quantum field theory and develops a methodology for deriving knot invariants from path integrals. We discuss challenges in quantizing metric independent actions and evaluating Wilson loop operators. Our proposed approach derives closed form expressions for linking numbers and demonstrates how perturbative expansions recover known knot invariants. Numerical examples illustrate how different gauge groups and levels affect knot polynomial values. The outcomes highlight the versatility of topological methods in physics and provide insight into potential applications in quantum computation and condensed matter. Our results show that simple mathematical constructs, when interpreted through gauge theory, yield powerful tools for classifying knots and three manifolds.",
      "url": "https://www.semanticscholar.org/paper/d1bbb01e1a465fbc460fae7f518d4b9663e2726c",
      "year": 2025,
      "venue": "Journal of Interdisciplinary Mathematics",
      "source": "semantic_scholar",
      "doi": "10.47974/jim-2370",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 571
    },
    {
      "title": "Spectral Theory and Hardy Spaces for Bessel Operators in Non-Standard Geometries",
      "authors": [
        "Saeed Hashemi Sababe"
      ],
      "abstract": "This paper develops novel results in the harmonic analysis of Bessel operators, extending their theory to higher-dimensional and non-Euclidean spaces. We present a refined framework for Hardy spaces associated with Bessel operators, emphasizing atomic decompositions, dual spaces, and connections to Sobolev and Besov spaces. The spectral theory of families of boundary-interpolating operators is also expanded, offering precise eigenvalue estimates and functional calculus applications. Furthermore, we explore Bessel operators under non-standard measures, such as fractal and weighted geometries, uncovering new analytical phenomena. Key implications include advanced insights into singular integrals, heat kernel behavior, and the boundedness of Riesz transforms, with potential applications in fractal geometry, constrained wave propagation, and mathematical physics.",
      "url": "https://www.semanticscholar.org/paper/0ae18009a2423ceb87eecc8f9ac06a55b7c51d49",
      "year": 2025,
      "venue": "Mathematics",
      "source": "semantic_scholar",
      "doi": "10.3390/math13040565",
      "pdf_url": "https://doi.org/10.3390/math13040565",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 572
    },
    {
      "title": "A non-Newtonian approach to electromagnetic curves in optical fiber",
      "authors": [
        "Aykut Has",
        "B. Yılmaz"
      ],
      "abstract": "The investigation within this article delves into the non-Newtonian geometric attributes exhibited by a linearly polarized light wave along an optical fiber within the framework of the 3D multiplicative Riemann manifold, employing multiplicative derivative and integral. While conducting this research, the unique arguments of multiplicative analysis (angle, norm, distance, etc.) are used. Within this context, the optical fiber is presumed as a one-dimensional entity embedded in the 3D Riemannian space, establishing a connection between the linearly polarized light wave's evolution and the geometric phase. Consequently, a novel form of the multiplicative geometric phase model is formulated, integrating the principles of multiplicative calculus. Additionally, the concept of multiplicative magnetic curves generated by the electric field $\\e$ is introduced. Notably, this study stands out due to its unique utilization of multiplicative derivatives and integrals in the computational processes. The article culminates by presenting illustrative examples consistent with the outlined theoretical framework, accompanied by visual representations. The distinctiveness of this research lies in its departure from conventional methodologies, incorporating multiplicative calculus into the calculations. Remarkably, multiplicative computing demonstrates its applicability across diverse domains including physics, engineering, mathematical biology, fluid mechanics, and signal processing. The pervasive use of multiplicative derivatives and integrals signifies their profound significance as a novel mathematical approach, contributing substantially to problem-solving methodologies across various scientific disciplines.",
      "url": "https://www.semanticscholar.org/paper/307e77baf0694e4e9a54e4c7b0f1d734c83099de",
      "year": 2025,
      "venue": "Revista mexicana de física",
      "source": "semantic_scholar",
      "doi": "10.31349/revmexfis.71.051306",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 573
    },
    {
      "title": "Segmentation of the spacecraft transfer problem through overdetermined and continuity constraints based on the Theory of Functional Connections",
      "authors": [
        "Allan Kardec de Almeida Junior"
      ],
      "abstract": "This paper introduces a segmented approach for solving constrained orbit transfer problems. The segments are connected through continuity constraints under the Theory of Functional Connections (TFC) mathematical framework that performs linear functional interpolation. This approach is further enhanced by a general vector formulation, from where the constrained functional is derived considering also a set of linear overdetermined constraints. Since we constrain vector instead of coordinates, this methodology allows to apply TFC to multiple and complex constraints composed by any types of nonlinear components included. We demonstrate the effectiveness of the method on Earth-to-Moon transfers, showing that this segmented approach achieves solutions with several orders of magnitude greater accuracy and efficiency in comparison with unsegmented orbit transfers.",
      "url": "https://www.semanticscholar.org/paper/947b30c6b9bf943b82a0cbc6a73bb97fdcfd2fe8",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 574
    },
    {
      "title": "High-order expansion of Neural Ordinary Differential Equations flows",
      "authors": [
        "Dario Izzo",
        "Sebastien Origer",
        "Giacomo Acciarini",
        "F. Biscani"
      ],
      "abstract": "Artificial neural networks, widely recognised for their role in machine learning, are now transforming the study of ordinary differential equations (ODEs), bridging data-driven modelling with classical dynamical systems and enabling the development of infinitely deep neural models. However, the practical applicability of these models remains constrained by the opacity of their learned dynamics, which operate as black-box systems with limited explainability, thereby hindering trust in their deployment. Existing approaches for the analysis of these dynamical systems are predominantly restricted to first-order gradient information due to computational constraints, thereby limiting the depth of achievable insight. Here, we introduce Event Transition Tensors, a framework based on high-order differentials that provides a rigorous mathematical description of neural ODE dynamics on event manifolds. We demonstrate its versatility across diverse applications: characterising uncertainties in a data-driven prey-predator control model, analysing neural optimal feedback dynamics, and mapping landing trajectories in a three-body neural Hamiltonian system. In all cases, our method enhances the interpretability and rigour of neural ODEs by expressing their behaviour through explicit mathematical structures. Our findings contribute to a deeper theoretical foundation for event-triggered neural differential equations and provide a mathematical construct for explaining complex system dynamics.",
      "url": "https://www.semanticscholar.org/paper/46e60e8ea9946c5aaacbe678b33b545b3229ecea",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2504.08769",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 575
    },
    {
      "title": "Dual Affine Connections, Legendre Transforms, and Black Hole Thermodynamics",
      "authors": [
        "Shoshauna Gauvin"
      ],
      "abstract": "Geometrical methods have become increasingly important in understanding both thermodynamics and information theory. In particular, dual affine (Hessian) geometry offers a powerful unification of concepts by recasting Legendre transformations as coordinate changes on a manifold endowed with a strictly convex potential. This viewpoint illuminates the mathematical basis of key thermodynamic relations, such as the mappings between internal energy U(S,V) and other potentials like Helmholtz or Gibbs free energies, and connects these ideas to the broader framework of information geometry, where dual coordinate systems naturally arise. In this paper, we present a concise treatment of how dual affine connections $(\\nabla, \\nabla^*)$ emerge from a single convex potential and are directly related through Legendre transforms. This emphasizes their physical significance and the geometric interpretation of entropy maximization. We then explore an energy gap integral constructed from the cubic form of the Hessian metric that measures how far a system deviates from the Levi Civita connection a self dual connection, and discuss how quantum-scale effects may render this gap infinite below the Planck length. Finally, we apply these concepts to black hole thermodynamics, showing how quantum or measurement uncertainties in (T,S,F,U) can be incorporated into the Hessian framework and interpreted via Hawking radiation in a stable black hole scenario. This unifying perspective underscores the natural extension from classical Riemannian geometry to dual-affine thermodynamics, with potential ramifications for quantum gravity and advanced thermodynamic modeling.",
      "url": "https://www.semanticscholar.org/paper/9a634935dfd147476f1fda8ddc84afdc5c276ac6",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 576
    },
    {
      "title": "Solving the Problem of Poor Internet Connectivity in Dhaka: Innovative Solutions Using Advanced WebRTC and Adaptive Streaming Technologies",
      "authors": [
        "Pavel Malinovskiy"
      ],
      "abstract": "Dhaka, Bangladesh, one of the world's most densely populated cities, faces severe challenges in maintaining reliable, high-speed internet connectivity. This paper presents an innovative framework that addresses poor mobile data connections through the integration of advanced WebRTC technology with adaptive streaming and server-side recording solutions. Focusing on the unique network conditions in Dhaka in 2025, our approach combines dynamic transcoding, real-time error correction, and optimized interface selection to enhance connectivity. We analyze empirical data on connection speeds, mobile tower density, district-level population statistics, and social media usage. Extensive mathematical formulations, including novel models for bitrate estimation, round-trip time optimization, and reliability analysis, are provided alongside detailed diagrams and multiple examples of code in both Python and C++. Experimental results demonstrate significant improvements in throughput, latency reduction, and overall service quality, offering a scalable blueprint for next-generation communication systems in hyper-dense urban environments.",
      "url": "https://www.semanticscholar.org/paper/51bfae27a69adda7e0659a96af622eb029b496e4",
      "year": 2025,
      "venue": "International Research Journal of Modernization in Engineering Technology and Science",
      "source": "semantic_scholar",
      "doi": "10.56726/IRJMETS68451",
      "pdf_url": "",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 577
    },
    {
      "title": "Communication-Efficient Device Scheduling for Federated Learning Using Lyapunov Optimization",
      "authors": [
        "Jake B. Perazzone",
        "Shiqiang Wang",
        "Mingyue Ji",
        "Kevin S. Chan"
      ],
      "abstract": "Federated learning (FL) is a useful tool that enables the training of machine learning models over distributed data without having to collect data centrally. When deploying FL in constrained wireless environments, however, intermittent connectivity of devices, heterogeneous connection quality, and non-i.i.d. data can severely slow convergence. In this paper, we consider FL with arbitrary device participation probabilities for each round and show that by weighing each device’s update by the reciprocal of their per-round participation probability, we can guarantee convergence to a stationary point. Our bound applies to non-convex loss functions and non-i.i.d. datasets and recovers state-of-the-art convergence rates for both full and uniform partial participation, including linear speedup, with only a single-sided learning rate. Then, using the derived convergence bound, we develop a new online client selection and power allocation algorithm that utilizes the Lyapunov drift-plus-penalty framework to opportunistically minimize a function of the convergence bound and the average communication time under a transmit power constraint. We use optimization over manifold techniques to obtain a solution to the minimization problem. Thanks to the Lyapunov framework, one key feature of the algorithm is that knowledge of the channel distribution is not required and only the instantaneous channel state information needs to be known. Using the CIFAR-10 dataset with varying levels of data heterogeneity, we show through simulations that the communication time can be significantly decreased using our algorithm compared to uniformly random participation, especially for heterogeneous channel conditions.",
      "url": "https://www.semanticscholar.org/paper/f85432c40947f00f91765424fdad77c79a437b81",
      "year": 2025,
      "venue": "IEEE Transactions on Networking",
      "source": "semantic_scholar",
      "doi": "10.1109/TON.2025.3539857",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 578
    },
    {
      "title": "Extension Research of Principal Bundle Constraint System Theory on Ricci-flat K\\\"ahler Manifolds",
      "authors": [
        "Dongzhe Zheng"
      ],
      "abstract": "This paper extends the geometric mechanics theory of constraint systems on principal bundles from the flat connection case to the general situation with non-zero curvature. Based on the theoretical foundation of compatible pairs under strong transversality conditions and principal bundle constraint systems, we systematically study the behavior of compatible pair theory in non-flat geometry within the context of Ricci-flat K\\\"ahler manifolds. Through rigorous mathematical analysis, we prove that the fundamental framework of strong transversality conditions and compatible pairs possesses universality and does not depend on the flatness assumption of connections. Furthermore, we re-derive the dynamical connection equations incorporating curvature reaction terms and extend Spencer cohomology theory to a spectral sequence structure capable of precisely encoding curvature information. The research reveals the exact action mechanism of curvature $\\Omega$: it plays a crucial role through the integrability condition $\\text{ad}_\\Omega^* \\lambda = 0$ and higher-order differentials in Spencer cohomology, while maintaining the fundamental structure of the theoretical framework unchanged. This extension not only deepens the geometric coupling theory between constraint systems and gauge fields but also provides new mathematical tools for modern gauge field theory, string theory geometry, and related topological analysis, demonstrating the profound connections between constraint geometry and classical differential geometry.",
      "url": "https://www.semanticscholar.org/paper/dd5ce52c0b3ac696e64f2c2e86910a715ca3c97a",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 579
    },
    {
      "title": "ODNet: Opinion Dynamics-Inspired Neural Message Passing for Graphs and Hypergraphs",
      "authors": [
        "Bingxin Zhou",
        "Outongyi Lv",
        "Jing Wang",
        "Xiang Xiao",
        "Weishu Zhao"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/3eef868a35888f318ca84f9810fa639ab324a63b",
      "year": 2025,
      "venue": "Trans. Mach. Learn. Res.",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 580
    },
    {
      "title": "The Study on Modified Theories of General Relativity: A Differential Geometric Approach",
      "authors": [
        "N. S. Kavya"
      ],
      "abstract": "The introduction of General Relativity (GR) in 1915 revolutionized our understanding of gravity, but over time, its limitations in explaining phenomena like dark energy, dark matter, and quantum gravity have motivated alternative theories. Early modifications, such as Weyl's 1919 proposal, focused on adding higher-order terms to the Einstein-Hilbert action. GR's non-renormalizability further strengthened the case for extending it. A central theme of modern gravity research is modifying the geometric structure, often by changing the gravitational Lagrangian. This leads to theories such as teleparallel and symmetric teleparallel gravity, utilizing torsion or non-Levi-Civita connections, with differential geometry providing the essential framework. This thesis explores several modified gravity models. Chapter 1 introduces necessary mathematical tools. Chapter 2 develops a novel parametrization of the deceleration parameter, constrained using MCMC and observational data, and applies it to f(Q) gravity. Chapter 3 embeds the LambdaCDM model into f(Q, L\\_m) gravity with non-minimal coupling, producing analytic solutions and matching observations through cosmographic analysis. Chapter 4 considers Bianchi-I spacetime in f(R, L\\_m) gravity with observational constraints to measure anisotropy. Chapter 5 presents wormhole solutions in f(Q, T) gravity with conformal symmetries. Chapter 6 explores wormholes in f(R, L\\_m) with non-commutative geometry, analyzing shape functions, energy conditions, and stability. Chapter 7 studies Big Bang Nucleosynthesis in f(T) gravity, constraining hybrid models using early- and late-time data, and validating intermediate epochs via cosmography.",
      "url": "https://www.semanticscholar.org/paper/05bab74866e77d18734622c03ac29eec4825f47e",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 581
    },
    {
      "title": "From exchangeability to rational belief: a cognitive interpretation of de Finetti’s theorem",
      "authors": [
        "Tommaso Costa"
      ],
      "abstract": "Probabilistic reasoning is central to many theories of human cognition, yet its foundations are often presented through abstract mathematical formalisms disconnected from the logic of belief and learning. In this article, we propose a reinterpretation of de Finetti’s representation theorem as a principle of rational inference under uncertainty. Building on the framework developed by E. T. Jaynes—where probability is viewed as an extension of logic—we show that the structure of de Finetti’s theorem mirrors the logic of belief updating constrained by symmetry. Exchangeable sequences, which treat observations as order-invariant, lead naturally to a representation in which probabilities are weighted averages over latent causes. This structure is formally analogous to the role of partition functions in statistical models, where uncertainty is distributed across hypotheses according to constraints and prior expectations. We argue that this correspondence is not merely mathematical but reveals a deeper cognitive interpretation: the mind, when faced with symmetry and incomplete information, may infer in ways that implicitly reflect maximum entropy principles. We illustrate this connection with a simple example and discuss how the underlying structure of de Finetti’s theorem can inform our understanding of inductive learning, probabilistic belief, and the rational architecture of cognition.",
      "url": "https://www.semanticscholar.org/paper/22d003260508e4243e89e0b258ebeca6b8639f52",
      "year": 2025,
      "venue": "Frontiers in Psychology",
      "source": "semantic_scholar",
      "doi": "10.3389/fpsyg.2025.1621552",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 582
    },
    {
      "title": "Measurability and continuity of parametric low-rank approximation in Hilbert spaces: linear operators and random variables",
      "authors": [
        "Nicola Rares Franco"
      ],
      "abstract": "\n We present a unified theoretical framework for parametric low-rank approximation, a research area devoted to the development of efficient algorithms that act as adaptive alternatives of traditional methods such as Singular Value Decomposition (SVD), Proper Orthogonal Decomposition (POD), and Principal Component Analysis (PCA). Applications include, e.g., the numerical treatment of parameter-dependent partial differential equations, where operators vary with parameters, and the statistical analysis of longitudinal data, where complex measurements, like audio signals and images, are collected over time. Recently, several adaptive algorithms have emerged, but a common mathematical foundation is still lacking, and existing solutions remain constrained to specific applications. As a result, key theoretical questions –such as the existence and regularity of optimal parametric low-rank approximants –remain inadequately addressed. Our goal is to bridge this gap between theory and practice by establishing a rigorous framework for parametric low-rank approximation under minimal assumptions, specifically focusing on cases where parameterizations are either measurable or continuous. The analysis is carried out within the context of separable Hilbert spaces, ensuring applicability to both finite and infinite-dimensional settings. Finally, connections to recently emerging trends in the Deep Learning literature, relevant for engineering and data science, are also discussed.",
      "url": "https://www.semanticscholar.org/paper/ac96b268db28eb03c5a03ec2719d6eb89992a305",
      "year": 2024,
      "venue": "Revista Matemática Complutense",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2409.09102",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 583
    },
    {
      "title": "Algorithmic and Machine Learning Methods for Witten Genus Vanishing Theorems",
      "authors": [
        "Song Xu"
      ],
      "abstract": "Witten genus, a fundamental invariant in index theory, extends $A$-genus and plays a crucial role in differential topology and mathematical physics. A central problem concerns identifying conditions under which the Witten genus vanishes, with the Stolz conjecture proposing a connection to positive Ricci curvature. While classical vanishing results exist for specific classes of complete intersections, there is no systematic computational approach for generating and analyzing such manifolds. This work introduces an algorithmic framework for studying the vanishing properties of the Witten genus in string complete intersections. By encoding complete intersections as configuration matrices, we develop an efficient algorithm for generating string manifolds in direct products of projective spaces and Bott manifolds and computing their (mod 2) Witten genera as q-series. As an application, we implement a predictive model using a convolutional neural network, achieving 0.9757 accuracy. CNN consistently outperforms logistic regression, random forests, and XGBoost across most training ratios, demonstrating that structural patterns associated with vanishing behavior can be effectively captured through machine learning.",
      "url": "https://www.semanticscholar.org/paper/d8c33cf02956e22c9154b54265b2c522dfb97ba8",
      "year": 2025,
      "venue": "2025 7th International Conference on Software Engineering and Computer Science (CSECS)",
      "source": "semantic_scholar",
      "doi": "10.1109/CSECS64665.2025.11009407",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 584
    },
    {
      "title": "Random Relay Jamming in Cooperative Free Space Optical Systems",
      "authors": [
        "Pratiti Paul",
        "M. Bhatnagar"
      ],
      "abstract": "Relay-assisted free space optical (FSO) systems offer several advantages such as cooperative diversity, mitigating turbulence-induced fading, and broad coverage area. Despite these manifold advantages, they can still be affected by unfavorable activities of jammer. Therefore, it is of high importance and need to investigate random jamming specially for security-constrained cooperative FSO communication protocol. This article develops a new system model of relay-assisted FSO system and studies a novel mathematical framework to analyze the performance of decode-and-forward protocol-based cooperative FSO system in presence of random relay jamming effects. In presence of jammer, the probability density functions (pdfs) of different additive noises, i.e., a mixture of negative exponential and Gaussian random variables (RVs) and mixture of negative exponential RVs are derived. A closed-form expression of the average bit error rate (ABER) is analytically evaluated by using the derived pdfs in the presence of random relay jamming effects. In presence of jammer, an approximate and simple expression of a threshold-based detector is obtained. By studying the ABER performance comparison, it is shown that the proposed approximate detector works satisfactorily compared to the maximum-likelihood detector in presence of random relay jammers. In addition to that, ABER performance of a general relay FSO system with an arbitrary number of relays is also obtained.",
      "url": "https://www.semanticscholar.org/paper/b4ca1b09d9fbcafd21830e5024a45ace376139cb",
      "year": 2022,
      "venue": "IEEE Systems Journal",
      "source": "semantic_scholar",
      "doi": "10.1109/jsyst.2021.3099019",
      "pdf_url": "",
      "citations": 11,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 585
    },
    {
      "title": "Orbit transfer using Theory of Functional Connections via change of variables",
      "authors": [
        "Allan K. de Almeida",
        "Antonio F. B. A. Prado",
        "D. Mortari"
      ],
      "abstract": "This work shows that a class of astrodynamics problems subject to mission constraints can be efficiently solved using the Theory of Functional Connections (TFC) mathematical framework by a specific change of coordinates. In these problems, the constraints are initially written in nonlinear and coupled mathematical forms using classical rectangular coordinates. The symmetries of the constrained problem are used to select a new system of coordinates that transforms the nonlinear constraints into linear. This change of coordinates is also used to isolate the components of the constraints. This way the TFC technique can be used to solve the ordinary differential equations governing orbit transfer problems subject to mission constraints. Specifically, this paper shows how to apply the change of coordinates method to the perturbed Hohmann-type and the one-tangent burn transfer problems.",
      "url": "https://www.semanticscholar.org/paper/73b9ac3572d2ea7434106e375817c532bfce2c96",
      "year": 2023,
      "venue": "The European Physical Journal Special Topics",
      "source": "semantic_scholar",
      "doi": "10.1140/epjs/s11734-023-01013-1",
      "pdf_url": "",
      "citations": 8,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 586
    },
    {
      "title": "Traveling waves in a model for cortical spreading depolarization with slow-fast dynamics.",
      "authors": [
        "David Reyner-Parra",
        "C. Bonet",
        "T. M. Seara",
        "G. Huguet"
      ],
      "abstract": "Cortical spreading depression and spreading depolarization (CSD) are waves of neuronal depolarization that spread across the cortex, leading to a temporary saturation of brain activity. They are associated with various brain disorders such as migraine and ischemia. We consider a reduced version of a biophysical model of a neuron-astrocyte network for the initiation and propagation of CSD waves [Huguet et al., Biophys. J. 111(2), 452-462, 2016], consisting of reaction-diffusion equations. The reduced model considers only the dynamics of the neuronal and astrocytic membrane potentials and the extracellular potassium concentration, capturing the instigation process implicated in such waves. We present a computational and mathematical framework based on the parameterization method and singular perturbation theory to provide semi-analytical results on the existence of a wave solution and to compute it jointly with its velocity of propagation. The traveling wave solution can be seen as a heteroclinic connection of an associated system of ordinary differential equations with a slow-fast dynamics. The presence of distinct time scales within the system introduces numerical instabilities, which we successfully address through the identification of significant invariant manifolds and the implementation of the parameterization method. Our results provide a methodology that allows to identify efficiently and accurately the mechanisms responsible for the initiation of these waves and the wave propagation velocity.",
      "url": "https://www.semanticscholar.org/paper/59c57d2a615005a99b1bae6d4d52b55f0c842488",
      "year": 2023,
      "venue": "Chaos",
      "source": "semantic_scholar",
      "doi": "10.1063/5.0160509",
      "pdf_url": "https://pubs.aip.org/aip/cha/article-pdf/doi/10.1063/5.0160509/18626232/083154_1_5.0160509.pdf",
      "citations": 4,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 587
    },
    {
      "title": "End-to-End Optimization of Metasurfaces for Imaging with Compressed Sensing",
      "authors": [
        "G. Arya",
        "William F. Li",
        "C. Roques-Carmes",
        "M. Soljačić",
        "Steven G. Johnson",
        "Zin Lin"
      ],
      "abstract": "We present a framework for the end-to-end optimization of metasurface imaging systems that reconstruct targets using compressed sensing, a technique for solving underdetermined imaging problems when the target object exhibits sparsity (i.e. the object can be described by a small number of non-zero values, but the positions of these values are unknown). We nest an iterative, unapproximated compressed sensing reconstruction algorithm into our end-to-end optimization pipeline, resulting in an interpretable, data-efficient method for maximally leveraging metaoptics to exploit object sparsity. We apply our framework to super-resolution imaging and high-resolution depth imaging with a phase-change material. In both situations, our end-to-end framework computationally discovers optimal metasurface structures for compressed sensing recovery, automatically balancing a number of complicated design considerations to select an imaging measurement matrix from a complex, physically constrained manifold with millions ofdimensions. The optimized metasurface imaging systems are robust to noise, significantly improving over random scattering surfaces and approaching the ideal compressed sensing performance of a Gaussian matrix, showing how a physical metasurface system can demonstrably approach the mathematical limits of compressed sensing.",
      "url": "https://www.semanticscholar.org/paper/8b5a821e2c87653e0ec4df88548b6c75ab5f21bc",
      "year": 2022,
      "venue": "ACS Photonics",
      "source": "semantic_scholar",
      "doi": "10.1021/acsphotonics.4c00259",
      "pdf_url": "https://arxiv.org/pdf/2201.12348",
      "citations": 27,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 588
    },
    {
      "title": "Discovering functional connectivity features characterizing multiple sclerosis phenotypes using explainable artificial intelligence",
      "authors": [
        "M. A. Yamin",
        "P. Valsasina",
        "J. Tessadori",
        "M. Filippi",
        "V. Murino",
        "M. Rocca",
        "Diego Sona"
      ],
      "abstract": "Multiple sclerosis (MS) is a neurological condition characterized by severe structural brain damage and by functional reorganization of the main brain networks that try to limit the clinical consequences of structural burden. Resting‐state (RS) functional connectivity (FC) abnormalities found in this condition were shown to be variable across different MS phases, according to the severity of clinical manifestations. The article describes a system exploiting machine learning on RS FC matrices to discriminate different MS phenotypes and to identify relevant functional connections for MS stage characterization. To this end, the system exploits some mathematical properties of covariance‐based RS FC representation, which can be described by a Riemannian manifold. The classification performance of the proposed framework was significantly above the chance level for all MS phenotypes. Moreover, the proposed system was successful in identifying relevant RS FC alterations contributing to an accurate phenotype classification.",
      "url": "https://www.semanticscholar.org/paper/e496c9dfa716f014507ed8e03e801346104ff2b6",
      "year": 2023,
      "venue": "Human Brain Mapping",
      "source": "semantic_scholar",
      "doi": "10.1002/hbm.26210",
      "pdf_url": "https://onlinelibrary.wiley.com/doi/pdfdirect/10.1002/hbm.26210",
      "citations": 6,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 589
    },
    {
      "title": "Hybrid geometrodynamics: a Hamiltonian description of classical gravity coupled to quantum matter",
      "authors": [
        "Jose Luis Alonso Buj",
        "Carlos Bouthelier Madre",
        "J. Clemente-Gallardo",
        "David Martínez-Crespo"
      ],
      "abstract": "We generalize the Hamiltonian picture of general relativity coupled to classical matter, known as geometrodynamics, to the case where such matter is described by a quantum field theory in curved spacetime, but gravity is still described by a classical metric tensor field over a spatial hypersurface and its associated momentum. Thus, in our approach there is no non-dynamic background structure, apart from the manifold of events, and the gravitational and quantum degrees of freedom have their dynamics inextricably coupled. Given the Hamiltonian nature of the framework, we work with the generators of hypersurface deformations over the manifold of quantum states. The construction relies heavily on the differential geometry of a fibration of the set of quantum states over the set of gravitational variables. An important mathematical feature of this work is the use of Minlos’s theorem to characterize Gaussian measures over the space of matter fields and of Hida distributions to define a common superspace to all possible Hilbert spaces with different measures, to properly characterize the Schrödinger wave functional picture of QFT in curved spacetime. This allows us to relate states within different Hilbert spaces in the case of vacuum states or measures that depend on the gravitational degrees of freedom, as the ones associated to Ashtekar’s complex structure. This is achieved through the inclusion of a quantum Hermitian connection for the fibration, which will have profound physical implications. The most remarkable physical features of the construction are norm conservation of the quantum state (even if the total dynamics are non-unitary), the clear identification of the hybrid conserved quantities and the description of a dynamical backreaction of quantum matter on geometry and vice versa, which shall modify the physical properties the gravitational field would have in the absence of backreaction.",
      "url": "https://www.semanticscholar.org/paper/fc9d2bd9b8e61a00a64c67fa241d2fb17e5f889d",
      "year": 2023,
      "venue": "Classical and quantum gravity",
      "source": "semantic_scholar",
      "doi": "10.1088/1361-6382/ad3459",
      "pdf_url": "https://arxiv.org/pdf/2307.00922",
      "citations": 7,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 590
    },
    {
      "title": "Topological indices of general relativity and Yang-Mills theory in four-dimensional space-time",
      "authors": [
        "Y. Kurihara"
      ],
      "abstract": "This report investigates general relativity and the Yang-Mills theory in four-dimensional space-time using a common mathematical framework, the Chern-Weil theory for principal bundles. The whole theory is described owing to the fibre bundle with the GL(4) symmetry by twisting several principal bundles with the gauge symmetry. In addition to the principal connection, we introduce the Hodge-dual connection into the Lagrangian to make gauge fields have dynamics independent from the Bianchi identity. We show that the duplex superstructure appears in the bundle when a Z2-grading operator exists in the total space of the bundle in general. The Dirac operator appears in the secondary superspace using the one-dimensional Clifford algebra, and it provides topological indices from the Atiyah-Singer index theorem. Though the topological index is usually discussed in the elliptic-type manifold, this report treats it in the hyperbolic-type space-time manifold using a novel method, the theta-metric space. The theta-metric treats the Euclidean and Minkowski spaces simultaneously and defines the topological index in the Minkowski space-time.",
      "url": "https://www.semanticscholar.org/paper/780aaf3ccf064741913efeaf7f41243cb6ca9c39",
      "year": 2022,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 2,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 591
    },
    {
      "title": "Stepsize anything: A unified learning rate schedule for budgeted-iteration training",
      "authors": [
        "Anda Tang",
        "Yiming Dong",
        "Yutao Zeng",
        "zhou Xun",
        "Zhouchen Lin"
      ],
      "abstract": "The expanding computational costs and limited resources underscore the critical need for budgeted-iteration training, which aims to achieve optimal learning within predetermined iteration budgets. While learning rate schedules fundamentally govern the performance of different networks and tasks, particularly in budgeted-iteration scenarios, their design remains largely heuristic, lacking theoretical foundations. In addition, the optimal learning rate schedule requires extensive trial-and-error selection, making the training process inefficient. In this work, we propose the Unified Budget-Aware (UBA) schedule, a theoretically grounded learning rate schedule that consistently outperforms commonly-used schedules among diverse architectures and tasks under different constrained training budgets. First, we bridge the gap by constructing a novel training budget-aware optimization framework, which explicitly accounts for the robustness to landscape curvature variations. From this framework, we derive the UBA schedule, controlled by a single hyper-parameter \\varphi that provides a trade-off between flexibility and simplicity, eliminating the need for per-network numerical optimization. Moreover, we establish a theoretical connection between \\varphi and the condition number, adding interpretation and justification to our approach. Besides, we prove the convergence for different values of \\varphi. We offer practical guidelines for its selection via theoretical analysis and empirical results. Extensive experimental results show that UBA consistently surpasses the commonly-used schedules across diverse vision and language tasks, spanning network architectures (e.g., ResNet, OLMo) and scales, under different training-iteration budgets.",
      "url": "https://www.semanticscholar.org/paper/5252fc9c44d42587da74fc166e18038fe8ee0263",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2505.24452",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 592
    },
    {
      "title": "DigiLoCS: A leap forward in predictive organ-on-chip simulations",
      "authors": [
        "M. R. Aravindakshan",
        "C. Mandal",
        "A. Pothen",
        "C. Maass"
      ],
      "abstract": "Digital twins, driven by data and mathematical modelling, have emerged as powerful tools for simulating complex biological systems. In this work, we focus on modelling the clearance on a liver-on-chip as a digital twin that closely mimics the clearance functionality of the human liver. Our approach involves the creation of a compartmental physiological model of the liver using ordinary differential equations (ODEs) to estimate pharmacokinetic (PK) parameters related to on-chip liver clearance. The objectives of this study were twofold: first, to predict human clearance values, and second, to propose a framework for bridging the gap between in vitro findings and their clinical relevance. The methodology integrated quantitative Organ-on-Chip (OoC) and cell-based assay analyses of drug depletion kinetics and is further enhanced by incorporating an OoC-digital twin model to simulate drug depletion kinetics in humans. The in vitro liver clearance for 32 drugs was predicted using a digital-twin model of the liver-on-chip and in vitro to in vivo extrapolation (IVIVE) was assessed using time series PK data. Three ODEs in the model define the drug concentrations in media, interstitium and intracellular compartments based on biological, hardware, and physicochemical information. A key issue in determining liver clearance appears to be the insufficient drug concentration within the intracellular compartment. The digital twin establishes a connection between the hardware chip structure and an advanced mapping of the underlying biology, specifically focusing on the intracellular compartment. Our modelling offers the following benefits: i) better prediction of intrinsic liver clearance of drugs compared to the state-of-the-art model and ii) explainability of behaviour based on physiological parameters. Finally, we illustrate the clinical significance of this approach by applying the findings to humans, utilising propranolol as a proof-of-concept example. This study stands out as the biggest cross-organ-on-chip platform investigation to date, systematically analysing and predicting human clearance values using data obtained from various in vitro liver-on-chip systems. Author summary Accurate prediction of in vivo clearance from in vitro data is important as inadequate understanding of the clearance of a compound can lead to unexpected and undesirable outcomes in clinical trials, ranging from underdosing to toxicity. Physiologically based pharmacokinetic (PBPK) model estimation of liver clearance is explored. The aim is to develop digital twins capable of determining better predictions of clinical outcomes, ultimately reducing the time, cost, and patient burden associated with drug development. Various hepatic in vitro systems are compared and their effectiveness for predicting human clearance is investigated. The developed tool, DigiLoCs, focuses explicitly on accurately describing complex biological processes within liver-chip systems. ODE-constrained optimisation is applied to estimate the clearance of compounds. DigiLoCs enable differentiation between active biological processes (metabolism) and passive processes (permeability and partitioning) by incorporating detailed information on compound-specific characteristics and hardware-specific data. These findings signify a significant stride towards more accurate and efficient drug development methodologies.",
      "url": "https://www.semanticscholar.org/paper/8a848c4ba71a6947b76050835050141c7827e4ff",
      "year": 2024,
      "venue": "bioRxiv",
      "source": "semantic_scholar",
      "doi": "10.1101/2024.03.28.587123",
      "pdf_url": "https://www.biorxiv.org/content/biorxiv/early/2024/03/29/2024.03.28.587123.full.pdf",
      "citations": 8,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 593
    },
    {
      "title": "OCTANE - Optimal Control for Tensor-based Autoencoder Network Emergence: Explicit Case",
      "authors": [
        "R. Khatri",
        "Anthony Kolshorn",
        "Colin Olson",
        "Harbir Antil"
      ],
      "abstract": "This paper presents a novel, mathematically rigorous framework for autoencoder-type deep neural networks that combines optimal control theory and low-rank tensor methods to yield memory-efficient training and automated architecture discovery. The learning task is formulated as an optimization problem constrained by differential equations representing the encoder and decoder components of the network and the corresponding optimality conditions are derived via a Lagrangian approach. Efficient memory compression is enabled by approximating differential equation solutions on low-rank tensor manifolds using an adaptive explicit integration scheme. These concepts are combined to form OCTANE (Optimal Control for Tensor-based Autoencoder Network Emergence) -- a unified training framework that yields compact autoencoder architectures, reduces memory usage, and enables effective learning, even with limited training data. The framework's utility is illustrated with application to image denoising and deblurring tasks and recommendations regarding governing hyperparameters are provided.",
      "url": "https://www.semanticscholar.org/paper/612607d5003ccec7cda114f93454e89cea1585d1",
      "year": 2025,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2509.08169",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 594
    },
    {
      "title": "Geometrical Modelling and Numerical Analysis of Dislocaion Mechanics",
      "authors": [
        "Shunsuke Kobayashi",
        "R. Tarumi"
      ],
      "abstract": "This study undertakes the mathematical modelling and numerical analysis of dislocations within the framework of differential geometry. The fundamental configurations, i.e. reference, intermediate and current configurations, are expressed as the Riemann-Cartan manifold, which equips the Riemannian metric and Weitzenb\\\"ock connection. The torsion 2-form on the intermediate configuration is obtained through the Hodge duality of the dislocation density and the corresponding bundle isomorphism is subjected to the Helmholtz decomposition. This analysis introduces the boundary condition for plastic deformation. Cartan first structure equation and stress equilibrium equation are solved numerically using weak form variational expressions and isogeometric analysis. The numerical analysis carried out for this study reveals the distribution of plastic deformation fields around screw and edge dislocations for the first time. It also demonstrates stress fields around dislocations of which the distant fields show full agreement with the classical Volterra theory, while at the same time eliminating the singularity otherwise introduced at the dislocation by classical methods. The stress fields include several characteristic features due to the geometrical nonlinearity included therein. We also demonstrate that free surfaces affect both plastic and elastic deformation, but in different ways. The mathematical framework of this study is applicable to an arbitrary configuration of dislocations.",
      "url": "https://www.semanticscholar.org/paper/06d4416902fe1137bd5c93b12ea53c0329088197",
      "year": 2022,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2205.02443",
      "pdf_url": "http://arxiv.org/pdf/2205.02443",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 595
    },
    {
      "title": "Quantifying the Structure of Disordered Materials",
      "authors": [
        "T. Hardin",
        "M. Chandross",
        "Rahul Meena",
        "Spencer Fajardo",
        "Dimitris G. Giovanis",
        "I. Kevrekidis",
        "M. Falk",
        "Michael Shields"
      ],
      "abstract": "Durable interest in developing a framework for the detailed structure of glassy materials has produced numerous structural descriptors that trade off between general applicability and interpretability. However, none approach the combination of simplicity and wide-ranging predictive power of the lattice-grain-defect framework for crystalline materials. Working from the hypothesis that the local atomic environments of a glassy material are constrained by enthalpy minimization to a low-dimensional manifold in atomic coordinate space, we develop a novel generalized distance function, the Gaussian Integral Inner Product (GIIP) distance, in connection with agglomerative clustering and diffusion maps, to parameterize that manifold. Applying this approach to a two-dimensional model crystal and a three-dimensional binary model metallic glass results in parameters interpretable as coordination number, composition, volumetric strain, and local symmetry. In particular, we show that a more slowly quenched glass has a higher degree of local tetrahedral symmetry at the expense of cyclic symmetry. While these descriptors require post-hoc interpretation, they minimize bias rooted in crystalline materials science and illuminate a range of structural trends that might otherwise be missed.",
      "url": "https://www.semanticscholar.org/paper/10cb93831ff385c2d101ee4ad4390f58fe1f551c",
      "year": 2022,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 596
    },
    {
      "title": "Rate-induced tipping: thresholds, edge states and connecting orbits",
      "authors": [
        "Sebastian Wieczorek",
        "Chunping Xie",
        "P. Ashwin"
      ],
      "abstract": "Rate-induced tipping (R-tipping) occurs when time-variation of input parameters of a dynamical system interacts with system timescales to give genuine nonautonomous instabilities. Such instabilities appear as the input varies at some critical rates and cannot, in general, be understood in terms of autonomous bifurcations in the frozen system with a fixed-in-time input. This paper develops an accessible mathematical framework for R-tipping in multidimensional nonautonomous dynamical systems with an autonomous future limit. We focus on R-tipping via loss of tracking of base attractors that are equilibria in the frozen system, due to crossing what we call regular R-tipping thresholds. These thresholds are anchored at infinity by regular R-tipping edge states: compact normally hyperbolic invariant sets of the autonomous future limit system that have one unstable direction, orientable stable manifold, and lie on a basin boundary. We define R-tipping and critical rates for the nonautonomous system in terms of special solutions that limit to a compact invariant set of the autonomous future limit system that is not an attractor. We focus on the case when the limit set is a regular edge state, introduce the concept of edge tails, and rigorously classify R-tipping into reversible, irreversible, and degenerate cases. The central idea is to use the autonomous dynamics of the future limit system to analyse R-tipping in the nonautonomous system. We compactify the original nonautonomous system to include the limiting autonomous dynamics. Considering regular R-tipping edge states that are equilibria allows us to prove two results. First, we give sufficient conditions for the occurrence of R-tipping in terms of easily testable properties of the frozen system and input variation. Second, we give necessary and sufficient conditions for the occurrence of reversible and irreversible R-tipping in terms of computationally verifiable (heteroclinic) connections to regular R-tipping edge states in the autonomous compactified system.",
      "url": "https://www.semanticscholar.org/paper/c5c5e699b11a269df38f1b1ea5822a9903f385ca",
      "year": 2021,
      "venue": "Nonlinearity",
      "source": "semantic_scholar",
      "doi": "10.1088/1361-6544/accb37",
      "pdf_url": "https://iopscience.iop.org/article/10.1088/1361-6544/accb37/pdf",
      "citations": 47,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 597
    },
    {
      "title": "A Battery-connected Switched-Capacitor-based Power Step-Up Converter for V2G Applications",
      "authors": [
        "Hakan Tekin",
        "Göknur Setrekli",
        "Eren Murtulu",
        "Hikmet Karşıyaka",
        "Davut Ertekin"
      ],
      "abstract": "In this study, a novel power boost converter topology, which is characterized by a single-switch configuration and its connection to a battery, is introduced. The primary objective of this study is to explore its suitability for voltage enhancement in electrification transportation systems. An intrinsic advantage of the proposed converter lies in its ability to provide a continuous current output with a constrained magnitude, consequently reducing peak values on the battery input side current as a critical parameter impacting battery reliability and longevity. Furthermore, the gain achieved by this novel converter is substantial, rendering it a viable choice for high-voltage Vehicle-to-Grid (V2G) applications.Unlike traditional boost converters, which usually produce an output voltage twice that of the input voltage with a duty cycle of 0.5, the suggested converter exceeds this by providing an output voltage five times higher than the input voltage. The validity of the theoretical framework is substantiated through mathematical derivations and simulation outcomes. To control the proposed converter effectively, a fuzzy logic controller is employed, chosen for its suitability in managing nonlinear systems and its simplicity of implementation.",
      "url": "https://www.semanticscholar.org/paper/30eac2064d9a636922707f64fc677d51e09e7962",
      "year": 2023,
      "venue": "International Conference on Electrical and Electronics Engineering",
      "source": "semantic_scholar",
      "doi": "10.1109/ELECO60389.2023.10416060",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 598
    },
    {
      "title": "Hojman-type conserved quantities for time-scale nonshifted mechanical systems",
      "authors": [
        "Shuang Hou",
        "Chuanjing Song"
      ],
      "abstract": "Conserved quantity is one of the core topics in analytical mechanics. Identifying such quantities in a system not only reduces the number of degrees of freedom, thereby simplifying the dynamical description, but also reveals the invariant laws that the system follows in the complex evolution process. In traditional dynamics, classifying system evolution as continuous or discrete serves as the foundation for building effective models and selecting appropriate mathematical tools. The time-scale theory, however, offers a unified framework for describing both types of systems. This approach not only avoids redundant derivations for each case but also reveals the profound structural connections between continuous and discrete dynamics. Consequently, exploring conserved quantities within the time-scale framework has emerged as a significant and valuable research direction. Unlike previous studies limited to Lagrangian system, this work establishes Hojman-type conserved quantities for time-scale nonshifted systems under both the Birkhoffian framework (including generalized and constrained Birkhoffian systems) and the Hamiltonian framework (covering both holonomic and nonholonomic cases). This study introduces, for the first time, a unified formulation for Lie symmetry and Hojman-type conserved quantity, with its application to both frameworks presented separately. All theoretical results are rigorously derived and supported by numerical simulations.",
      "url": "https://www.semanticscholar.org/paper/ccf2247db3ac168ea3f4762bf99abd4740e0d99e",
      "year": 2025,
      "venue": "Physica Scripta",
      "source": "semantic_scholar",
      "doi": "10.1088/1402-4896/ae2cff",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 599
    },
    {
      "title": "Robot-assisted mapping of chemical reaction hyperspaces and networks",
      "authors": [
        "Yankai Jia",
        "Rafał Frydrych",
        "Yaroslav I. Sobolev",
        "Wai-Shing Wong",
        "Bibek Prajapati",
        "Daniel Matuszczyk",
        "Yasemin Bilgi",
        "Louis Gadina",
        "Juan Carlos Ahumada",
        "Galymzhan Moldagulov"
      ],
      "abstract": "Despite decades of investigation, it remains unclear (and hard to predict1, 2, 3–4) how the outcomes of chemical reactions change over multidimensional ‘hyperspaces’ defined by reaction conditions5. Whereas human chemists can explore only a limited subset of these manifolds, automated platforms6, 7, 8, 9, 10, 11–12 can generate thousands of reactions in parallel. Yet, purification and yield quantification remain bottlenecks, constrained by time-consuming and resource-intensive analytical techniques. As a result, our understanding of reaction hyperspaces remains fragmentary7,9,13, 14, 15–16. Are yield distributions smooth or corrugated? Do they conceal mechanistically new reactions? Can major products vary across different regions? Here, to address these questions, we developed a low-cost robotic platform using primarily optical detection to quantify yields of products and by-products at unprecedented throughput and minimal cost per condition. Scanning hyperspaces across thousands of conditions, we find and prove mathematically that, for continuous variables (concentrations, temperatures), individual yield distributions are generally slow-varying. At the same time, we uncover hyperspace regions of unexpected reactivity as well as switchovers between major products. Moreover, by systematically surveying substrate proportions, we reconstruct underlying reaction networks and expose hidden intermediates and products—even in reactions studied for well over a century. This hyperspace-scanning approach provides a versatile and scalable framework for reaction optimization and discovery. Crucially, it can help identify conditions under which complex mixtures can be driven cleanly towards different major products, thereby expanding synthetic diversity while reducing chemical input requirements. A low-cost robotic platform using mainly optical detection to quantify yields of products and by-products allows the analysis of multidimensional chemical reaction hyperspaces and networks much faster than is possible by human chemists.",
      "url": "https://www.semanticscholar.org/paper/db54e93e27f6702c48cb182bf529da4a4c97836b",
      "year": 2025,
      "venue": "Nature",
      "source": "semantic_scholar",
      "doi": "10.1038/s41586-025-09490-1",
      "pdf_url": "",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 600
    },
    {
      "title": "THE ABSTRACTS OF THE TALKS 2022 ALGEBRA AND BEYOND A CONFERENCE IN HONOR OF THE MATHEMATICAL CONTRIBUTIONS OF MICHAEL J. LARSEN",
      "authors": [
        "Chun Yin Hui",
        "C. Simpson",
        "A. Lindenstrauss",
        "M. Wood",
        "Shekhar Khare",
        "C. Chai"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/57338a5c8efc977ae75eda4fd1421c2c745aa758",
      "year": 2022,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 601
    },
    {
      "title": "Preparing Code States via Seed-Entangler-Enriched Sequential Quantum Circuits: Application to Tetra-Digit Topological Error-Correcting Codes",
      "authors": [
        "Yu-Tao Hu",
        "Meng-Yuan Li",
        "Peng Ye"
      ],
      "abstract": "Demonstrating how long-range entangled states are born from product states has gained much attention, which is not only important for quantum technology but also provides an unconventional tool in characterizing and classifying exotic phases of matter. In this paper, we introduce a unified and efficient framework of quantum circuits (i.e., a series of local unitary transformations), termed the \\emph{Seed-Entangler-Enriched Sequential Quantum Circuit} (SEESQC) to construct long-range entangled states (i.e., code states) in code space of topological error-correcting codes. Specifically, we apply SEESQC to construct code states of Tetra-Digit models -- a broad class of long-range entangled stabilizer codes indexed by a four-digit parameter. These models are not rare but encompass Toric Codes across arbitrary dimensions and subsume the X-cube fracton code as special cases. Featuring a hierarchical structure of generalized entanglement renormalization group, many Tetra-Digit models host spatially extended excitations (e.g., loops, membranes, and exotic non-manifold objects) with constrained mobility and deformability, and exhibit system-size-dependent ground state degeneracies that scale exponentially with a polynomial in linear sizes. In this work, we begin with graphical and algebraic demonstration of quantum circuits for computational basis states, before generalizing to broader cases. Central to this framework is a key ingredient termed the \\emph{seed-entangler} acting on a small number of qubits termed \\textit{seeds}, enabling a systematic scheme to achieve arbitrary code states. Remarkably, the number of available seeds equals the number of logical qubits for the constructed examples, which leaves plenty of room for future investigation in theoretical physics, mathematics and quantum information science. Beyond the critical limitation of prior state-engineering methodologies, ...",
      "url": "https://www.semanticscholar.org/paper/f3293f326e4221f842754a62691666239e8e8fee",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": "10.1103/d8gs-fnwt",
      "pdf_url": "",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 602
    },
    {
      "title": "Combinatorial Cell Complexes: Duality, reconstruction and causal cobordisms",
      "authors": [
        "Maxime Savoy"
      ],
      "abstract": "This thesis proposes a framework based on a notion of combinatorial cell complex (cc) whose cells are defined simply as finite sets of vertices. The cells of a cc are subject to four axioms involving a rank function that assigns a rank (or a dimension) to each cell. Our framework focuses on classes of cc admitting an inclusion-reversing duality map. We introduce a combinatorial notion of cobordism that allows us to single out a category whose morphisms are cobordisms having a causal structure. Our aim is to offer an approach to look for a combinatorial notion of quantum field theory having a built-in duality operation acting on the underlying space and not relying on any manifold structure. The introduction includes links with fields in Theoretical and Mathematical Physics related to Quantum Gravity and motivating our framework. We start by introducing cc and the duality map on a class of cc with empty boundary called closed cc. We then focus on the problem of reconstructing a certain class of cc from their cells of rank 2 and lower. Such cc are in particular duals to simplicial complexes with no boundary and their reconstruction is realized using a discrete notion of connection. Our next main result extends the duality map we defined on closed cc to a class of cc with boundary. An important by-product of the study of this extended duality map is the combinatorial notion of cobordism used in this work. We also introduce a general notion of subdivision of a cc via a map called reduction, as well as the dual notion of reduction called collapse. These two types of map characterize the structure of certain cc called slices, using sequences of maps called slice sequences. Slices are the basic building blocs of our definition of causal cobordisms and the dual of a slice sequence defines the composition of cobordisms, providing us with a category whose morphisms are causal cobordisms.",
      "url": "https://www.semanticscholar.org/paper/01ec853346b7a9ba0abf8d98a5907bcb119a538a",
      "year": 2022,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 603
    },
    {
      "title": "Mathematical programming formulations for piecewise polynomial functions",
      "authors": [
        "B. Grimstad",
        "B. Knudsen"
      ],
      "abstract": "This paper studies mathematical programming formulations for solving optimization problems with piecewise polynomial (PWP) constraints. We elaborate on suitable polynomial bases as a means of efficiently representing PWPs in mathematical programs, comparing and drawing connections between the monomial basis, the Bernstein basis, and B-splines. The theory is presented for both continuous and semi-continuous PWPs. Using a disjunctive formulation, we then exploit the characteristic of common polynomial basis functions to significantly reduce the number of nonlinearities, and to suggest a bound-tightening technique for PWP constraints. We derive several extensions using Bernstein cuts, an expanded Bernstein basis, and an expanded monomial basis, which upon a standard big-M reformulation yield a set of new MINLP models. The formulations are compared by globally solving six test sets of MINLPs and a realistic petroleum production optimization problem. The proposed framework shows promising numerical performance and facilitates the solution of PWP-constrained optimization problems using standard MINLP software.",
      "url": "https://www.semanticscholar.org/paper/67afb6ca1c18e78566e9eb220ca62ef93a9876e3",
      "year": 2020,
      "venue": "Journal of Global Optimization",
      "source": "semantic_scholar",
      "doi": "10.1007/s10898-020-00881-4",
      "pdf_url": "https://link.springer.com/content/pdf/10.1007/s10898-020-00881-4.pdf",
      "citations": 5,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 604
    },
    {
      "title": "Network controllability measures of subnetworks: implications for neurosciences",
      "authors": [
        "Jule E. Stocker",
        "Erfan Nozari",
        "Marieke K. van Vugt",
        "A. Jansen",
        "H. Jamalabadi"
      ],
      "abstract": "Objective: Recent progress in network sciences has made it possible to apply key findings from control theory to the study of networks. Referred to as network control theory, this framework describes how the interactions between interconnected system elements and external energy sources, potentially constrained by different optimality criteria, result in complex network behavior. A typical example is the quantification of the functional role certain brain regions or symptoms play in shaping the temporal dynamics of brain activity or the clinical course of a disease, a property that is quantified in terms of the so-called controllability metrics. Critically though, contrary to the engineering context in which control theory was originally developed, a mathematical understanding of the network nodes and connections in neurosciences cannot be assumed. For instance, in the case of psychological systems such as those studied to understand psychiatric disorders, a potentially large set of related variables are unknown. As such, while the measures offered by network control theory would be mathematically correct, in that they can be calculated with high precision, they could have little translational values with respect to their putative role suggested by controllability metrics. It is therefore critical to understand if and how the controllability metrics estimated over subnetworks would deviate, if access to the complete set of variables, as is common in neurosciences, cannot be taken for granted. Approach: In this paper, we use a host of simulations based on synthetic as well as structural magnetic resonance imaging (MRI) data to study the potential deviation of controllability metrics in sub- compared to the full networks. Specifically, we estimate average- and modal-controllability, two of the most widely used controllability measures in neurosciences, in a large number of settings where we systematically vary network type, network size, and edge density. Main results: We find out, across all network types we test, that average and modal controllability are systematically, over- or underestimated depending on the number of nodes in the sub- and full network and the edge density. Significance: Finally, we provide formal theoretical proof that our observations generalize to any network type and discuss the ramifications of this systematic bias and potential solutions to alleviate the problem.",
      "url": "https://www.semanticscholar.org/paper/4ebd644ffff1c15629755847822aec64a5ec92cc",
      "year": 2022,
      "venue": "bioRxiv",
      "source": "semantic_scholar",
      "doi": "10.1088/1741-2552/acb256",
      "pdf_url": "https://pure.rug.nl/ws/files/606451182/Stocker_2023_J._Neural_Eng._20_016044.pdf",
      "citations": 4,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 605
    },
    {
      "title": "Revisiting Volterra defects: geometrical relation between edge dislocations and wedge disclinations",
      "authors": [
        "Shunsuke Kobayashi",
        "Katsumi Takemasa",
        "R. Tarumi"
      ],
      "abstract": "This study presents a comprehensive mathematical model for Volterra defects and explores their relations using differential geometry on Riemann–Cartan manifolds. Following the standard Volterra process, we derived the Cartan moving frame, a geometric representation of plastic fields, and the associated Riemannian metric using exterior algebra. Although the analysis naturally defines the geometry of three types of dislocations and the wedge disclination, it fails to classify twist disclinations owing to the persistent torsion component, suggesting the need for modifications to the Volterra process. By leveraging the interchangeability of the Weitzenböck and Levi-Civita connections and applying an analytical solution for plasticity derived from the Biot–Savart law, we provide a rigorous mathematical proof of the long-standing phenomenological relationship between edge dislocations and wedge disclinations. Additionally, we showcase the effectiveness of novel mathematical tools, including Riemannian holonomy for analysing the Frank vector and complex potentials that encapsulate the topological properties of wedge disclinations as jump discontinuities. Furthermore, we derive analytical expressions for the linearized stress fields of wedge disclinations and confirm their consistency with existing results. These findings demonstrate that the present geometrical framework extends and generalizes the classical theory of Volterra defects.",
      "url": "https://www.semanticscholar.org/paper/d1245220b70419c82ab312c3ab0b79debf077298",
      "year": 2024,
      "venue": "Royal Society Open Science",
      "source": "semantic_scholar",
      "doi": "10.1098/rsos.242213",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 606
    },
    {
      "title": "An Open Unified Addressing System for 6G Communication Networks",
      "authors": [
        "Guanwen Li",
        "D. Lou",
        "A. Galis",
        "Jinze Yang",
        "Chuang Wang",
        "Sheng Jiang",
        "Zhe Chen",
        "Xing Tong"
      ],
      "abstract": "With the rapid and continuous development of the Internet, it is foreseeable that current addressing schemes and fixed-length IP addresses would create further bottlenecks and limitations in realizing future 6G networking requirements, such as massive connections, resource-constrained communication, and heterogeneous hyper interconnections and guaranteeing agreement-based services and KPIs. Moreover, the locator-based addressing semantic is unsuitable for mobile and content-oriented networks. Thus, this paper proposes the Open Unified Addressing (OUA) system, a novel, flexible, multi-semantic and hierarchical addressing architecture that better supports the flexibility and extensibility of the Internet protocol framework in the context of 6G Communications. The OUA addresses several limitations in the current IP protocol and improves communication efficiency. According to the evaluation with two typical forwarding models, the results show that the OUA system has almost no impact on forwarding delay. Moreover, it can provide scalable addressing spaces and shorten the route convergence time.",
      "url": "https://www.semanticscholar.org/paper/e052320b7d61e70e69f249f3052527051709a1ae",
      "year": 2022,
      "venue": "2022 IEEE Future Networks World Forum (FNWF)",
      "source": "semantic_scholar",
      "doi": "10.1109/FNWF55208.2022.00034",
      "pdf_url": "https://discovery.ucl.ac.uk/10166968/1/OUA_FNWF_WS7.pdf",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 607
    },
    {
      "title": "Computing Topological Indices and Polynomials of the Rhenium Trioxide",
      "authors": [
        "S. Imran",
        "Muhammad Mudassar Raza",
        "N. Nigar",
        "S. Kirmani",
        "F. B. Petros"
      ],
      "abstract": "In the study of mathematical chemistry and chemical graph theory, a topological index, also known as a connectivity index, is the arithmetical framework of a graph that specifies its topology and also graph invariant. These topological indices are used to model quantitative structure relationships \n \n \n \n Q\n S\n A\n R\n s\n \n \n \n , which are connections between the work of biological or other molecular structures and the chemical structures. This study computed the first, second, and Hyper Zagreb indices, as well as Zagreb polynomials, Redefined Zagreb indices, Randic index, \n \n A\n B\n C\n \n index, and \n \n G\n A\n \n index of chemical structure of Rhenium Trioxide.",
      "url": "https://www.semanticscholar.org/paper/7736643e6caab370dcb1b7a3c11180870647e5a4",
      "year": 2022,
      "venue": "Journal of mathematics",
      "source": "semantic_scholar",
      "doi": "10.1155/2022/4838327",
      "pdf_url": "https://downloads.hindawi.com/journals/jmath/2022/4838327.pdf",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 608
    },
    {
      "title": "Integrated Multi-Objective Optimization for Reheating Furnace Scheduling and Rolling Plan in Hot Rolling Process of Steel Industry",
      "authors": [
        "Qi Wang",
        "Zhongyang Han",
        "Jun Zhao",
        "Wei Wang"
      ],
      "abstract": "As one of the most important process, hot rolling that consists of several procedures exhibits as both a primary production period and an intensive energy-consuming consumer, of which the production efficiency and energy consumption is of great significance for the enterprises. Although existed methods have proposed some solutions, most of them are mainly based on single-objective optimization toward separate procedure rather than multi-objective alternative considering the practical connections between the consecutive stages. For this purpose, an integrated optimization method based on unified modeling and multi-stage evolutionary algorithm is proposed in this study. More specifically, a mathematical model for this Constrained Multi-Objective Problem (CMOP) is formulated at first, where the objectives and constraints for the scheduling of reheating furnace and rolling is ensemble considered. Furthermore, an improved Multi-Stage Multi-Objective Evolutionary Algorithm framework (MSMOEA) is proposed, which performs a sequential Pareto Front based optimization according to the flow of production procedures, so that a scheduling plan for both the reheating furnace and rolling is finally obtained. To evaluate the performance for both the modeling and optimization, the proposed unified model and MSMOEA framework are compared with single-objective model and commonly deployed optimization methods using real data from a steel plant in China, respectively. The results demonstrate the superiority of the proposed method, which can be beneficial for efficiency increasement and energy costs reduction on site.",
      "url": "https://www.semanticscholar.org/paper/cff9df082579596d6037c1f8d252098d086eddd1",
      "year": 2022,
      "venue": "Chinese Control and Decision Conference",
      "source": "semantic_scholar",
      "doi": "10.1109/CCDC55256.2022.10033616",
      "pdf_url": "",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 609
    },
    {
      "title": "Aharonov-Bohm Effects for Electromagnetism and Gravity in Four-Dimensional Spacetime",
      "authors": [
        "Yanhui Li",
        "Y. Reyimuaji"
      ],
      "abstract": "This paper investigates a geometric framework for the gravitational Aharonov-Bohm effect in four-dimensional spacetime, demonstrating how spacetime curvature induces nonlocal quantum phase shifts within field-free regions. By constructing vector bundles on spacetime manifolds equipped with Levi-Civita connections, we derive the holonomy transformations for parallel-transported quantum states. Under the Newtonian approximation, metric decomposition into Minkowski background plus scalar potential perturbations reveals through the linearized Einstein field equations that the gravitationally induced phase shift is mathematically isomorphic to its electromagnetic counterpart. These results establish the quantum observability of gravitational gauge structures and provide theoretical support for experimental verification via atom interferometry.",
      "url": "https://www.semanticscholar.org/paper/75a4adba8e0d9bfff240eb167b97e05d2a47408a",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 610
    },
    {
      "title": "Vulcan: Instance-Optimal Systems Heuristics Through LLM-Driven Search",
      "authors": [
        "Rohit Dwivedula",
        "Divyanshu Saxena",
        "Sujay Yadalam",
        "Daehyeok Kim",
        "Aditya Akella"
      ],
      "abstract": "Resource-management tasks in modern operating and distributed systems continue to rely primarily on hand-designed heuristics for tasks such as scheduling, caching, or active queue management. Designing performant heuristics is an expensive, time-consuming process that we are forced to continuously go through due to the constant flux of hardware, workloads and environments.\n  We propose a new alternative: synthesizing instance-optimal heuristics -- specialized for the exact workloads and hardware where they will be deployed -- using code-generating large language models (LLMs). To make this synthesis tractable, Vulcan separates policy and mechanism through LLM-friendly, task-agnostic interfaces. With these interfaces, users specify the inputs and objectives of their desired policy, while Vulcan searches for performant policies via evolutionary search over LLM-generated code. This interface is expressive enough to capture a wide range of system policies, yet sufficiently constrained to allow even small, inexpensive LLMs to generate correct and executable code.\n  We use Vulcan to synthesize performant heuristics for cache eviction and memory tiering, and find that these heuristics outperform all human-designed state-of-the-art algorithms by upto 69% and 7.9% in performance for each of these tasks respectively.",
      "url": "http://arxiv.org/abs/2512.25065v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25065v1",
      "citations": null,
      "categories": [
        "cs.OS",
        "cs.AI",
        "cs.DC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 611
    },
    {
      "title": "Feeling Blue: Constructing a Robust SALT3 UV Template and Constraining its Redshift Dependency",
      "authors": [
        "Qinan Wang",
        "David O. Jones",
        "Justin D. R. Pierel",
        "Matthew R. Siebert",
        "W. D'Arcy Kenworthy",
        "Richard Kessler",
        "Mi Dai",
        "Ryan J. Foley",
        "Ori D. Fox",
        "Suvi Gezari"
      ],
      "abstract": "Upcoming cosmological surveys will obtain numerous rest-frame ultraviolet (UV) observations of Type Ia supernovae (SNe Ia), yet there is concern about how standardizable SNe Ia are in the UV. In this work, we train a robust optical--UV SED model for SNe Ia (SALT3-UV) with the open-source model-training software $\\texttt{SALTshaker}$. We incorporate a spectroscopic UV data sample from HST, including 67 UV spectra from 18 nearby SNe Ia. Unlike previous training spectra, the HST spectra have sufficiently precise calibration that they do not require additional warping to match coincident photometric data. Additionally, while including this new SN Ia sample necessitates incorporating auxiliary photometric data from ZTF and ATLAS that have insufficient calibration for cosmological analyses, the improvements in the calibration of these data is anticipated in the near future. Compared to the previous SALT3-K21 model, the SALT3-UV model shows a significant improvement in the UV down to $2000\\mathring{\\text{A}}$, with over a threefold improvement in model uncertainty and a more physically accurate continuum and line features. We further evaluate potential redshift evolution in the UV template by separating the UV training sample into low- and high-$z$ subsamples. Our results reveal a non-negligible $\\gtrsim 0.05$ mag difference between low- and high-$z$ SALT3-UV models in the $g-$band at $z\\gtrsim0.5$ and the $u-$band at $z\\gtrsim0.2$. We demonstrate that, if confirmed, such evolution could lead to a few-percent bias in the measurement of $w$ if high-$z$ rest-frame UV data are included in future cosmological surveys such as LSST and $\\textit{Roman}$.",
      "url": "http://arxiv.org/abs/2512.25064v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25064v1",
      "citations": null,
      "categories": [
        "astro-ph.CO",
        "astro-ph.HE"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 612
    },
    {
      "title": "Numerical study of solitary waves in Dirac--Klein--Gordon system",
      "authors": [
        "Andrew Comech",
        "Julien Ricaud",
        "Marco Roque"
      ],
      "abstract": "We use numerics to construct solitary waves in Dirac--Klein--Gordon (in one and three spatial dimensions) and study the dependence of energy and charge on $ω$. For the construction, we use the iterative procedure, starting from solitary waves of nonlinear Dirac equation, computing the corresponding scalar field, and adjusting the coupling constant. We also consider the case of massless scalar field, when the iteration procedure could be compared with the shooting method. We use the virial identities to control the error of simulations. We also discuss possible implications from the obtained results for the spectral stability of solitary waves.",
      "url": "http://arxiv.org/abs/2512.24954v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24954v1",
      "citations": null,
      "categories": [
        "math-ph",
        "math.AP"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 613
    },
    {
      "title": "Dynamic response phenotypes and model discrimination in systems and synthetic biology",
      "authors": [
        "Eduardo D. Sontag"
      ],
      "abstract": "Biological systems encode function not primarily in steady states, but in the structure of transient responses elicited by time-varying stimuli. Overshoots, biphasic dynamics, adaptation kinetics, fold-change detection, entrainment, and cumulative exposure effects often determine phenotypic outcomes, yet are poorly captured by classical steady-state or dose-response analyses. This paper develops an input-output perspective on such \"dynamic phenotypes,\" emphasizing how qualitative features of transient behavior constrain underlying network architectures independently of detailed parameter values. A central theme is the role of sign structure and interconnection logic, particularly the contrast between monotone systems and architectures containing antagonistic pathways. We show how incoherent feedforward (IFF) motifs provide a simple and recurrent mechanism for generating non-monotonic and adaptive responses across multiple levels of biological organization, from molecular signaling to immune regulation and population dynamics. Conversely, monotonicity imposes sharp impossibility results that can be used to falsify entire classes of models from transient data alone. Beyond step inputs, we highlight how periodic forcing, ramps, and integral-type readouts such as cumulative dose responses offer powerful experimental probes that reveal otherwise hidden structure, separate competing motifs, and expose invariances such as fold-change detection. Throughout, we illustrate how control-theoretic concepts, including monotonicity, equivariance, and input-output analysis, can be used not as engineering metaphors, but as precise mathematical tools for biological model discrimination. Thus we argue for a shift in emphasis from asymptotic behavior to transient and input-driven dynamics as a primary lens for understanding, testing, and reverse-engineering biological networks.",
      "url": "http://arxiv.org/abs/2512.24945v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24945v1",
      "citations": null,
      "categories": [
        "q-bio.QM",
        "math.OC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 614
    },
    {
      "title": "Securing High-Concurrency Ticket Sales: A Framework Based on Microservice",
      "authors": [
        "Zhiyong Zhang",
        "Xiaoyan Zhang",
        "Xiaoqi Li"
      ],
      "abstract": "The railway ticketing system is one of the most important public service infrastructure. In peak periods such as holidays, it is often faced with the challenge of high concurrency scenarios because of a large number of users accessing at the same time. The traditional aggregation architecture can not meet the peak user requirements because of its insufficient fault tolerance and low ability. Therefore, the system needs to use microservice architecture for development, and add multiple security methods to ensure that the system can have good stability and data consistency under high concurrency scenarios, and can respond quickly to user requests. This paper introduces the use of B/S architecture and Spring Cloud to design and develop a railway ticket purchase system that can maintain stability and reliability under high concurrency scenarios, and formulate multiple security design methods for the system. This system integrates a range of functions, such as real-time train inquiries, dynamic seat updates, online seat selection, and ticket purchasing, effectively addressing common problems associated with offline ticket purchasing, such as long queues and delayed information. It enables a complete online process from inquiry and booking to payment and refunds. Furthermore, the \"add passenger\" function allows users to purchase tickets for others, extending the convenience of online ticketing to people with limited internet access. The system design prioritizes security and stability, while also focusing on high performance, and achieves these goals through a carefully designed architecture and the integration of multiple middleware components. After the completion of the system development, the core interface of the system is tested, and then the results are analyzed. The test data proves that the system has good ability and stability under high concurrency.",
      "url": "http://arxiv.org/abs/2512.24941v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24941v1",
      "citations": null,
      "categories": [
        "cs.SE"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 615
    },
    {
      "title": "Constraints on the perfect phylogeny mixture model and their effect on reducing degeneracy",
      "authors": [
        "John Marangola",
        "Azadeh Sheikholeslami",
        "José Bento"
      ],
      "abstract": "The perfect phylogeny mixture (PPM) model is useful due to its simplicity and applicability in scenarios where mutations can be assumed to accumulate monotonically over time. It is the underlying model in many tools that have been used, for example, to infer phylogenetic trees for tumor evolution and reconstruction. Unfortunately, the PPM model gives rise to substantial ambiguity -- in that many different phylogenetic trees can explain the same observed data -- even in the idealized setting where data are observed perfectly, i.e. fully and without noise. This ambiguity has been studied in this perfect setting by Pradhan et al. 2018, which proposed a procedure to bound the number of solutions given a fixed instance of observation data. Beyond this, studies have been primarily empirical. Recent work (Myers et al. 2019) proposed adding extra constraints to the PPM model to tackle ambiguity. In this paper, we first show that the extra constraints of Myers et al. 2019, called longitudinal constraints (LC), often fail to reduce the number of distinct trees that explain the observations. We then propose novel alternative constraints to limit solution ambiguity and study their impact when the data are observed perfectly. Unlike the analysis in Pradhan et al. 2018, our theoretical results regarding both the inefficacy of the LC and the extent to which our new constrains reduce ambiguity are not tied to a single observation instance. Rather, our theorems hold over large ensembles of possible inference problems. To the best of our knowledge, we are the first to study degeneracy in the PPM model in this ensemble-based theoretical framework.",
      "url": "http://arxiv.org/abs/2512.24930v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24930v1",
      "citations": null,
      "categories": [
        "q-bio.PE",
        "stat.OT"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 616
    },
    {
      "title": "Cosmological dynamics and observational constraints of an interacting early scalar field coupled to radiation",
      "authors": [
        "Dorian Araya",
        "Felipe Herrera",
        "Nelson Videla"
      ],
      "abstract": "We study the cosmic evolution of an interacting scalar field radiation model, in which a minimally coupled scalar field exchanges energy with the radiation sector through an exponential coupling. Extending previous formulations, a non-relativistic matter component is included explicitly, which allows a self consistent description of cosmological dynamics from the radiation-dominated era to late-time acceleration. Analytical expressions for the background expansion are derived and characterized using kinematic diagnostics. We constrain the model using observational Hubble data, Type Ia Supernovae, baryon acoustic oscillations (including DESI DR2), and compressed cosmic microwave background distance information, performing a Bayesian MCMC analysis. The interaction parameter is found to be consistent with zero, though small deviations from standard radiation scaling are allowed. These deviations can partially alleviate the Hubble tension by modifying the sound horizon, but this is accompanied by correlated shifts in the matter density. The reconstructed expansion history remains close to LCDM at late times. Model comparison suggest that the interacting scenario is statistically competitive but not decisively preferred by current background data.",
      "url": "http://arxiv.org/abs/2512.24918v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24918v1",
      "citations": null,
      "categories": [
        "astro-ph.CO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 617
    },
    {
      "title": "Stochastic factors can matter: improving robust growth under ergodicity",
      "authors": [
        "Balint Binkert",
        "David Itkin",
        "Paul Mangers Bastian",
        "Josef Teichmann"
      ],
      "abstract": "Drifts of asset returns are notoriously difficult to model accurately and, yet, trading strategies obtained from portfolio optimization are very sensitive to them. To mitigate this well-known phenomenon we study robust growth-optimization in a high-dimensional incomplete market under drift uncertainty of the asset price process $X$, under an additional ergodicity assumption, which constrains but does not fully specify the drift in general. The class of admissible models allows $X$ to depend on a multivariate stochastic factor $Y$ and fixes (a) their joint volatility structure, (b) their long-term joint ergodic density and (c) the dynamics of the stochastic factor process $Y$. A principal motivation of this framework comes from pairs trading, where $X$ is the spread process and models with the above characteristics are commonplace. Our main results determine the robust optimal growth rate, construct a worst-case admissible model and characterize the robust growth-optimal strategy via a solution to a certain partial differential equation (PDE). We demonstrate that utilizing the stochastic factor leads to improvement in robust growth complementing the conclusions of the previous study by Itkin et. al. (arXiv:2211.15628 [q-fin.MF], forthcoming in $\\textit{Finance and Stochastics}$), which additionally robustified the dynamics of the stochastic factor leading to $Y$-independent optimal strategies. Our analysis leads to new financial insights, quantifying the improvement in growth the investor can achieve by optimally incorporating stochastic factors into their trading decisions. We illustrate our theoretical results on several numerical examples including an application to pairs trading.",
      "url": "http://arxiv.org/abs/2512.24906v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24906v1",
      "citations": null,
      "categories": [
        "q-fin.MF",
        "math.PR"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 618
    },
    {
      "title": "One-Shot Camera-Based Extrusion Optimization for High Speed Fused Filament Fabrication",
      "authors": [
        "Yufan Lin",
        "Xavier Guidetti",
        "Yannick Nagel",
        "Efe C. Balta",
        "John Lygeros"
      ],
      "abstract": "Off-the-shelf fused filament fabrication 3D printers are widely accessible and convenient, yet they exhibit quality loss at high speeds due to dynamic mis-synchronization between printhead motion and material extrusion systems, notably corner over-extrusion. Existing methods require specialized hardware, extensive calibration, or firmware modifications that are inaccessible to most users. This work presents a practical, end-to-end optimization framework that enhances high-speed printing using only standard 3D printers and a phone camera, without requiring additional complex setup. The method employs a one-shot calibration approach in which two simple printed patterns, captured by a phone camera, enable identification of extrusion dynamics and cornering behavior. The identified systems enable a model-based constrained optimal control strategy that generates optimized G-code, synchronizing motion and extrusion. Experiments show reduced width tracking error, mitigated corner defects, and lower surface roughness, achieving surface quality at 3600 mm/min comparable to conventional printing at 1600 mm/min, effectively doubling production speed while maintaining print quality. This accessible, hardware-minimal approach enables a wide range of fused filament fabrication users to achieve high-quality, high-speed additive manufacturing.",
      "url": "http://arxiv.org/abs/2512.24905v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24905v1",
      "citations": null,
      "categories": [
        "eess.SY"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 619
    },
    {
      "title": "MTSP-LDP: A Framework for Multi-Task Streaming Data Publication under Local Differential Privacy",
      "authors": [
        "Chang Liu",
        "Junzhou Zhao"
      ],
      "abstract": "The proliferation of streaming data analytics in data-driven applications raises critical privacy concerns, as directly collecting user data may compromise personal privacy. Although existing $w$-event local differential privacy (LDP) mechanisms provide formal guarantees without relying on trusted third parties, their practical deployment is hindered by two key limitations. First, these methods are designed primarily for publishing simple statistics at each timestamp, making them inherently unsuitable for complex queries. Second, they handle data at each timestamp independently, failing to capture temporal correlations and consequently degrading the overall utility. To address these issues, we propose MTSP-LDP, a novel framework for \\textbf{M}ulti-\\textbf{T}ask \\textbf{S}treaming data \\textbf{P}ublication under $w$-event LDP. MTSP-LDP adopts an \\emph{Optimal Privacy Budget Allocation} algorithm to dynamically allocate privacy budgets by analyzing temporal correlations within each window. It then constructs a \\emph{data-adaptive private binary tree structure} to support complex queries, which is further refined by cross-timestamp grouping and smoothing operations to enhance estimation accuracy. Furthermore, a unified \\emph{Budget-Free Multi-Task Processing} mechanism is introduced to support a variety of streaming queries without consuming additional privacy budget. Extensive experiments on real-world datasets demonstrate that MTSP-LDP consistently achieves high utility across various streaming tasks, significantly outperforming existing methods.",
      "url": "http://arxiv.org/abs/2512.24899v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24899v1",
      "citations": null,
      "categories": [
        "cs.CR"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 620
    },
    {
      "title": "Towards autonomous time-calibration of large quantum-dot devices: Detection, real-time feedback, and noise spectroscopy",
      "authors": [
        "Anantha S. Rao",
        "Barnaby van Straaten",
        "Valentin John",
        "Cécile X. Yu",
        "Stefan D. Oosterhout",
        "Lucas Stehouwer",
        "Giordano Scappucci",
        "M. D. Stewart,",
        "Menno Veldhorst",
        "Francesco Borsoi"
      ],
      "abstract": "The performance and scalability of semiconductor quantum-dot (QD) qubits are limited by electrostatic drift and charge noise that shift operating points and destabilize qubit parameters. As systems expand to large one- and two-dimensional arrays, manual recalibration becomes impractical, creating a need for autonomous stabilization frameworks. Here, we introduce a method that uses the full network of charge-transition lines in repeatedly acquired double-quantum-dot charge stability diagrams (CSDs) as a multidimensional probe of the local electrostatic environment. By accurately tracking the motion of selected transitions in time, we detect voltage drifts, identify abrupt charge reconfigurations, and apply compensating updates to maintain stable operating conditions. We demonstrate our approach on a 10-QD device, showing robust stabilization and real-time diagnostic access to dot-specific noise processes. The high acquisition rate of radio-frequency reflectometry CSD measurements also enables time-domain noise spectroscopy, allowing the extraction of noise power spectral densities, the identification of two-level fluctuators, and the analysis of spatial noise correlations across the array. From our analysis, we find that the background noise at 100~$μ$\\si{\\hertz} is dominated by drift with a power law of $1/f^2$, accompanied by a few dominant two-level fluctuators and an average linear correlation length of $(188 \\pm 38)$~\\si{\\nano\\meter} in the device. These capabilities form the basis of a scalable, autonomous calibration and characterization module for QD-based quantum processors, providing essential feedback for long-duration, high-fidelity qubit operations.",
      "url": "http://arxiv.org/abs/2512.24894v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24894v1",
      "citations": null,
      "categories": [
        "cond-mat.mes-hall",
        "cs.CV",
        "cs.ET",
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 621
    },
    {
      "title": "Adaptive Clutter Suppression via Convex Optimization",
      "authors": [
        "Yifan He",
        "Griffin Kearney",
        "Makan Fardad"
      ],
      "abstract": "Passive and bistatic radar systems are often limited by strong clutter and direct-path interference that mask weak moving targets. Conventional cancellation methods such as the extensive cancellation algorithm require careful tuning and can distort the delay-Doppler response. This paper introduces a convex optimization framework that adaptively synthesizes per-cell delay-Doppler filters to suppress clutter while preserving the canonical cross-ambiguity function (CAF). The approach formulates a quadratic program that minimizes distortion of the CAF surface subject to linear clutter-suppression constraints, eliminating the need for a separate cancellation stage. Monte Carlo simulations using common communication waveforms demonstrate strong clutter suppression, accurate CFAR calibration, and major detection-rate gains over the classical CAF. The results highlight a scalable, CAF-faithful method for adaptive clutter mitigation in passive radar.",
      "url": "http://arxiv.org/abs/2512.24889v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24889v1",
      "citations": null,
      "categories": [
        "math.OC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 622
    },
    {
      "title": "SoK: Web3 RegTech for Cryptocurrency VASP AML/CFT Compliance",
      "authors": [
        "Qian'ang Mao",
        "Jiaxin Wang",
        "Ya Liu",
        "Li Zhu",
        "Jiaman Chen",
        "Jiaqi Yan"
      ],
      "abstract": "The decentralized architecture of Web3 technologies creates fundamental challenges for Anti-Money Laundering and Counter-Financing of Terrorism compliance. Traditional regulatory technology solutions designed for centralized financial systems prove inadequate for blockchain's transparent yet pseudonymous networks. This systematization examines how blockchain-native RegTech solutions leverage distributed ledger properties to enable novel compliance capabilities.\n  We develop three taxonomies organizing the Web3 RegTech domain: a regulatory paradigm evolution framework across ten dimensions, a compliance protocol taxonomy encompassing five verification layers, and a RegTech lifecycle framework spanning preventive, real-time, and investigative phases. Through analysis of 41 operational commercial platforms and 28 academic prototypes selected from systematic literature review (2015-2025), we demonstrate that Web3 RegTech enables transaction graph analysis, real-time risk assessment, cross-chain analytics, and privacy-preserving verification approaches that are difficult to achieve or less commonly deployed in traditional centralized systems.\n  Our analysis reveals critical gaps between academic innovation and industry deployment, alongside persistent challenges in cross-chain tracking, DeFi interaction analysis, privacy protocol monitoring, and scalability. We synthesize architectural best practices and identify research directions addressing these gaps while respecting Web3's core principles of decentralization, transparency, and user sovereignty.",
      "url": "http://arxiv.org/abs/2512.24888v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24888v1",
      "citations": null,
      "categories": [
        "cs.CR",
        "cs.SE"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 623
    },
    {
      "title": "A structure-preserving parametric approximation for anisotropic geometric flows via an $α$-surface energy matrix",
      "authors": [
        "Weizhu Bao",
        "Yifei Li",
        "Wenjun Ying",
        "Yulin Zhang"
      ],
      "abstract": "We propose a structure-preserving parametric approximation for geometric flows with general anisotropic effects. By introducing a hyperparameter $α$, we construct a unified surface energy matrix $\\hat{\\boldsymbol{G}}_k^α(θ)$ that encompasses all existing formulations of surface energy matrices, and apply it to anisotropic curvature flow. We prove that $α=-1$ is the unique choice achieving optimal energy stability under the necessary and sufficient condition $3\\hatγ(θ)\\geq\\hatγ(θ-π)$, while all other $α\\neq-1$ require strictly stronger conditions. The framework extends naturally to general anisotropic geometric flows through a unified velocity discretization that ensures energy stability. Numerical experiments validate the theoretical optimality of $α=-1$ and demonstrate the effectiveness and robustness.",
      "url": "http://arxiv.org/abs/2512.24875v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24875v1",
      "citations": null,
      "categories": [
        "math.NA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 624
    },
    {
      "title": "Encyclo-K: Evaluating LLMs with Dynamically Composed Knowledge Statements",
      "authors": [
        "Yiming Liang",
        "Yizhi Li",
        "Yantao Du",
        "Ge Zhang",
        "Jiayi Zhou",
        "Yuchen Wu",
        "Yinzhu Piao",
        "Denghui Cao",
        "Tong Sun",
        "Ziniu Li"
      ],
      "abstract": "Benchmarks play a crucial role in tracking the rapid advancement of large language models (LLMs) and identifying their capability boundaries. However, existing benchmarks predominantly curate questions at the question level, suffering from three fundamental limitations: vulnerability to data contamination, restriction to single-knowledge-point assessment, and reliance on costly domain expert annotation. We propose Encyclo-K, a statement-based benchmark that rethinks benchmark construction from the ground up. Our key insight is that knowledge statements, not questions, can serve as the unit of curation, and questions can then be constructed from them. We extract standalone knowledge statements from authoritative textbooks and dynamically compose them into evaluation questions through random sampling at test time. This design directly addresses all three limitations: the combinatorial space is too vast to memorize, and model rankings remain stable across dynamically generated question sets, enabling reliable periodic dataset refresh; each question aggregates 8-10 statements for comprehensive multi-knowledge assessment; annotators only verify formatting compliance without requiring domain expertise, substantially reducing annotation costs. Experiments on over 50 LLMs demonstrate that Encyclo-K poses substantial challenges with strong discriminative power. Even the top-performing OpenAI-GPT-5.1 achieves only 62.07% accuracy, and model performance displays a clear gradient distribution--reasoning models span from 16.04% to 62.07%, while chat models range from 9.71% to 50.40%. These results validate the challenges introduced by dynamic evaluation and multi-statement comprehensive understanding. These findings establish Encyclo-K as a scalable framework for dynamic evaluation of LLMs' comprehensive understanding over multiple fine-grained disciplinary knowledge statements.",
      "url": "http://arxiv.org/abs/2512.24867v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24867v1",
      "citations": null,
      "categories": [
        "cs.CL",
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 625
    },
    {
      "title": "Latent Twins: A Framework for Scene Recognition and Fast Radiative Transfer Inversion in FORUM All-Sky Observations",
      "authors": [
        "Cristina Sgattoni",
        "Luca Sgheri",
        "Matthias Chung",
        "Michele Martinazzo"
      ],
      "abstract": "The FORUM (Far-infrared Outgoing Radiation Understanding and Monitoring) mission will provide, for the first time, systematic far-infrared spectral measurements of Earth's outgoing radiation, enabling improved understanding of atmospheric processes and the radiation budget. Retrieving atmospheric states from these observations constitutes a high-dimensional, ill-posed inverse problem, particularly under cloudy-sky conditions where multiple-scattering effects are present. In this work, we develop a data-driven, physics-aware inversion framework for FORUM all-sky retrievals based on latent twins: coupled autoencoders for atmospheric states and spectra, combined with bidirectional latent-space mappings. A lightweight model-consistency correction ensures physically plausible cloud variable reconstructions. The resulting framework demonstrates potential for retrievals of atmospheric, cloud and surface variables, providing information that can serve as a prior, initial guess, or surrogate for computationally expensive full-physics inversion methods. It also enables robust scene classification and near-instantaneous inference, making it suitable for operational near-real-time applications. We demonstrate its performance on synthetic FORUM-like data and discuss implications for future data assimilation and climate studies.",
      "url": "http://arxiv.org/abs/2512.24865v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24865v1",
      "citations": null,
      "categories": [
        "physics.ao-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 626
    },
    {
      "title": "Antecedents of Consumer Regret Frequency: The Roles of Decision Agency, Status Signaling, and Online Shopping Preference",
      "authors": [
        "Shawn Berry"
      ],
      "abstract": "Consumer regret is a widespread post-purchase emotion that significantly impacts satisfaction, product returns, complaint behavior, and customer loyalty. Despite its prevalence, there is a limited understanding of why certain consumers experience regret more frequently as a chronic aspect of their engagement in the marketplace. This study explores the antecedents of consumer regret frequency by integrating decision agency, status signaling motivations, and online shopping preferences into a cohesive framework. By analyzing survey data (n=338), we assess whether consumers' perceived agency and decision-making orientation correlate with the frequency of regret, and whether tendencies towards status-related consumption and preferences for online shopping environments exacerbate regret through mechanisms such as increased social comparison, expanded choice sets, and continuous exposure to alternative offers. The findings reveal that regret frequency is significantly linked to individual differences in decision-related orientations and status signaling, with a preference for online shopping further contributing to regret-prone consumption behaviors. These results extend the scope of regret and cognitive dissonance research beyond isolated decision episodes by emphasizing regret frequency as a persistent consumer outcome. From a managerial standpoint, the findings suggest that retailers can alleviate regret-driven dissatisfaction by enhancing decision support, minimizing choice overload, and developing post-purchase reassurance strategies tailored to segments prone to regret..",
      "url": "http://arxiv.org/abs/2512.24862v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24862v1",
      "citations": null,
      "categories": [
        "econ.GN"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 627
    },
    {
      "title": "OFL-SAM2: Prompt SAM2 with Online Few-shot Learner for Efficient Medical Image Segmentation",
      "authors": [
        "Meng Lan",
        "Lefei Zhang",
        "Xiaomeng Li"
      ],
      "abstract": "The Segment Anything Model 2 (SAM2) has demonstrated remarkable promptable visual segmentation capabilities in video data, showing potential for extension to medical image segmentation (MIS) tasks involving 3D volumes and temporally correlated 2D image sequences. However, adapting SAM2 to MIS presents several challenges, including the need for extensive annotated medical data for fine-tuning and high-quality manual prompts, which are both labor-intensive and require intervention from medical experts. To address these challenges, we introduce OFL-SAM2, a prompt-free SAM2 framework for label-efficient MIS. Our core idea is to leverage limited annotated samples to train a lightweight mapping network that captures medical knowledge and transforms generic image features into target features, thereby providing additional discriminative target representations for each frame and eliminating the need for manual prompts. Crucially, the mapping network supports online parameter update during inference, enhancing the model's generalization across test sequences. Technically, we introduce two key components: (1) an online few-shot learner that trains the mapping network to generate target features using limited data, and (2) an adaptive fusion module that dynamically integrates the target features with the memory-attention features generated by frozen SAM2, leading to accurate and robust target representation. Extensive experiments on three diverse MIS datasets demonstrate that OFL-SAM2 achieves state-of-the-art performance with limited training data.",
      "url": "http://arxiv.org/abs/2512.24861v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24861v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 628
    },
    {
      "title": "VLN-MME: Diagnosing MLLMs as Language-guided Visual Navigation agents",
      "authors": [
        "Xunyi Zhao",
        "Gengze Zhou",
        "Qi Wu"
      ],
      "abstract": "Multimodal Large Language Models (MLLMs) have demonstrated remarkable capabilities across a wide range of vision-language tasks. However, their performance as embodied agents, which requires multi-round dialogue spatial reasoning and sequential action prediction, needs further exploration. Our work investigates this potential in the context of Vision-and-Language Navigation (VLN) by introducing a unified and extensible evaluation framework to probe MLLMs as zero-shot agents by bridging traditional navigation datasets into a standardized benchmark, named VLN-MME. We simplify the evaluation with a highly modular and accessible design. This flexibility streamlines experiments, enabling structured comparisons and component-level ablations across diverse MLLM architectures, agent designs, and navigation tasks. Crucially, enabled by our framework, we observe that enhancing our baseline agent with Chain-of-Thought (CoT) reasoning and self-reflection leads to an unexpected performance decrease. This suggests MLLMs exhibit poor context awareness in embodied navigation tasks; although they can follow instructions and structure their output, their 3D spatial reasoning fidelity is low. VLN-MME lays the groundwork for systematic evaluation of general-purpose MLLMs in embodied navigation settings and reveals limitations in their sequential decision-making capabilities. We believe these findings offer crucial guidance for MLLM post-training as embodied agents.",
      "url": "http://arxiv.org/abs/2512.24851v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24851v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.RO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 629
    },
    {
      "title": "AODDiff: Probabilistic Reconstruction of Aerosol Optical Depth via Diffusion-based Bayesian Inference",
      "authors": [
        "Linhao Fan",
        "Hongqiang Fang",
        "Jingyang Dai",
        "Yong Jiang",
        "Qixing Zhang"
      ],
      "abstract": "High-quality reconstruction of Aerosol Optical Depth (AOD) fields is critical for Atmosphere monitoring, yet current models remain constrained by the scarcity of complete training data and a lack of uncertainty quantification.To address these limitations, we propose AODDiff, a probabilistic reconstruction framework based on diffusion-based Bayesian inference. By leveraging the learned spatiotemporal probability distribution of the AOD field as a generative prior, this framework can be flexibly adapted to various reconstruction tasks without requiring task-specific retraining. We first introduce a corruption-aware training strategy to learns a spatiotemporal AOD prior solely from naturally incomplete data. Subsequently, we employ a decoupled annealing posterior sampling strategy that enables the more effective and integration of heterogeneous observations as constraints to guide the generation process. We validate the proposed framework through extensive experiments on Reanalysis data. Results across downscaling and inpainting tasks confirm the efficacy and robustness of AODDiff, specifically demonstrating its advantage in maintaining high spatial spectral fidelity. Furthermore, as a generative model, AODDiff inherently enables uncertainty quantification via multiple sampling, offering critical confidence metrics for downstream applications.",
      "url": "http://arxiv.org/abs/2512.24847v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24847v1",
      "citations": null,
      "categories": [
        "cs.LG",
        "physics.ao-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 630
    },
    {
      "title": "All optical Lithography for Spatiotemporal Patterning of Azopolymer Microreliefs",
      "authors": [
        "I Komang Januariyasa",
        "Francesco Reda",
        "Nikolai Liubimtsev",
        "Marina Saphiannikova",
        "Fabio Borbone",
        "Marcella Salvatore",
        "Stefano Luigi Oscurato"
      ],
      "abstract": "Microstructured surfaces are central to photonics, biointerfaces, and functional coatings, yet they are typically fabricated through multi-step lithographic workflows requiring masks or molds and post-processing. Azopolymers provide an alternative route by converting structured optical fields into surface reliefs via light-induced mass migration, but existing approaches have been limited to smooth, shallow, and engraving-like topographies produced from a flat film. Here we introduce an all-optical, maskless, fully digital lithography platform that exploits engineered darkness within computer-generated holograms to spatially localize inward mass transport and directly produce positive, protruding microreliefs. We show that isolated and array of micro-bumps can be generated from pristine flat azopolymer films in a single writing step, and we introduce spatiotemporal control through sequential tailored illumination to reshape microrelief profiles, enabling flattened-top micropillars, programmable array shapes and arrangements, and free-form continuous microrelief designs. Hierarchical microarchitectures are also demonstrated by extending the concept of multi-step illumination sequences. As functional demonstrations, we realize multi-focus microlenses and quasi-square diffraction gratings with enhanced 1st-order efficiencies. Finally, we leverage azopolymer reconfigurability to implement write-erase-rewrite cycles that reset and repurpose the same surface region for distinct micropatterns, enabling rewritable surfaces and reprogrammable master templates for replication. Overall, this work establishes a scalable spatiotemporal strategy for on-demand, all-optical microfabrication and reprogramming of structured surfaces, where spatial and temporal degrees of freedom of holographic patterns intermix to produce advanced patterning capabilities.",
      "url": "http://arxiv.org/abs/2512.25048v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25048v1",
      "citations": null,
      "categories": [
        "physics.optics"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 631
    },
    {
      "title": "Universal polar dual pairs of spherical codes found in $E_8$ and $Λ_{24}$",
      "authors": [
        "S. V. Borodachov",
        "P. G. Boyvalenkov",
        "P. D. Dragnev",
        "D. P. Hardin",
        "E. B. Saff",
        "M. M. Stoyanova"
      ],
      "abstract": "We identify universal polar dual pairs of spherical codes $C$ and $D$ such that for a large class of potential functions $h$ the minima of the discrete $h$-potential of $C$ on the sphere occur at the points of $D$ and vice versa. Moreover, the minimal values of their normalized potentials are equal. These codes arise from the known sharp codes embedded in the even unimodular extremal lattices $E_8$ and $Λ_{24}$ (Leech lattice). This embedding allows us to use the lattices' properties to find new universal polar dual pairs. In the process we extensively utilize the interplay between the binary Golay codes and the Leech lattice.\n  As a byproduct of our analysis, we identify a new universally optimal (in the sense of energy) code in the projective space $\\mathbb{RP}^{21}$ with $1408$ points (lines). Furthermore, we extend the Delsarte-Goethals-Seidel definition of derived codes from their seminal $1977$ paper and generalize their Theorem 8.2 to show that if a $τ$-design is enclosed in $k\\leq τ$ parallel hyperplanes, then each of the hyperplane's sub-code is a $(τ+1-k)$-design in the ambient subspace.",
      "url": "http://arxiv.org/abs/2512.25037v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25037v1",
      "citations": null,
      "categories": [
        "math.CO",
        "math.CA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 632
    },
    {
      "title": "Strengthening Dual Bounds for Multicommodity Capacitated Network Design with Unsplittable Flow Constraints",
      "authors": [
        "Lacy M. Greening",
        "Santanu S. Dey",
        "Alan L. Erera"
      ],
      "abstract": "Multicommodity capacitated network design (MCND) models can be used to optimize the consolidation of shipments within e-commerce fulfillment networks. In practice, fulfillment networks require that shipments with the same origin and destination follow the same transfer path. This unsplittable flow requirement complicates the MCND problem, requiring integer programming (IP) formulations in which binary variables replace continuous flow variables. To enhance the solvability of this variant of the MCND problem for large-scale logistics networks, this work focuses on strengthening dual bounds. We investigate the polyhedra of arc-set relaxations, and we introduce two new classes of valid inequalities that can be implemented within solution approaches. We develop one approach that dynamically adds valid inequalities to the root node of a reformulation of the MCND IP with additional valid metric inequalities. We show the effectiveness of our ideas with a comprehensive computational study using path-based fulfillment instances, constructed from data provided by a large U.S.-based e-commerce company, and the well-known arc-based Canad instances. Experiments show that our best solution approach for a practical path-based model reduces the IP gap by an average of 26.5% and 22.5% for the two largest instance groups, compared to solving the reformulation alone, demonstrating its effectiveness in improving the dual bound. In addition, experiments using only the arc-based relaxation highlight the strength of our new valid inequalities relative to the linear programming relaxation (LPR), yielding an IP-gap reduction of more than 85%.",
      "url": "http://arxiv.org/abs/2512.25018v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25018v1",
      "citations": null,
      "categories": [
        "math.OC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 633
    },
    {
      "title": "Noise resilient real-time phase imaging via undetected light",
      "authors": [
        "Josué R. León-Torres",
        "Patrick Hendra",
        "Yugant Mukeshbhai Hadiyal",
        "Christopher Spiess",
        "Fabian Steinlechner",
        "Frank Setzpfandt",
        "Markus Gräfe",
        "Valerio Flavio Gili"
      ],
      "abstract": "Quantum imaging with undetected light has recently emerged as a technique in which quantum correlations and nonlinear interferometry are combined to decouple illumination and detection paths. This approach has been more recently extended and combined with digital phase-shifting holography and off-axis holography to extract both the amplitude and phase information of a sample relying on single-photon interference. Despite these advantages, implementing the technique in real-world scenarios where the observed system is subject to environmental noise and dynamic variations remains challenging. The primary limitation lies in the inability of quantum imaging systems to retrieve object information in real time under high-noise conditions. Here, we experimentally demonstrate real-time amplitude and phase imaging in noisy environments, building upon our previous implementation of quantum off-axis holography. Our results demonstrate real-time imaging at acquisition rates up to 4~Hz, even when the noise level exceeds the signal by an order of magnitude.",
      "url": "http://arxiv.org/abs/2512.24993v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24993v1",
      "citations": null,
      "categories": [
        "physics.optics"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 634
    },
    {
      "title": "A guide to the $2$-generated axial algebras of Monster type",
      "authors": [
        "Justin McInroy",
        "Abdul Wajid Mir"
      ],
      "abstract": "Axial algebras of Monster type are a class of non-associative algebras which generalise the Griess algebra, whose automorphism group is the largest sporadic simple group, the Monster. The $2$-generated algebras, which are the building blocks from which all algebras in this class can be constructed, have recently been classified by Yabe; Franchi and Mainardis; and Franchi, Mainardis and McInroy. There are twelve infinite families of examples as well as the exceptional Highwater algebra and its cover, however their properties are not well understood.\n  In this paper, we detail the properties of each of these families, describing their ideals and quotients, subalgebras and idempotents in all characteristics. We also describe all exceptional isomorphisms between them. We give new bases for several of the algebras which better exhibit their axial features and provide code for others to work with them.",
      "url": "http://arxiv.org/abs/2512.24987v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24987v1",
      "citations": null,
      "categories": [
        "math.RA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 635
    },
    {
      "title": "Any Clifford+T circuit can be controlled with constant T-depth overhead",
      "authors": [
        "Isaac H. Kim",
        "Tuomas Laakkonen"
      ],
      "abstract": "Since an n-qubit circuit consisting of CNOT gates can have up to $Ω(n^2/\\log{n})$ CNOT gates, it is natural to expect that $Ω(n^2/\\log{n})$ Toffoli gates are needed to apply a controlled version of such a circuit. We show that the Toffoli count can be reduced to at most n. The Toffoli depth can also be reduced to O(1), at the cost of 2n Toffoli gates, even without using any ancilla or measurement. In fact, using a measurement-based uncomputation, the Toffoli depth can be further reduced to 1. From this, we give two corollaries: any controlled Clifford circuit can be implemented with O(1) T-depth, and any Clifford+T circuit with T-depth D can be controlled with T-depth O(D), even without ancillas. As an application, we show how to catalyze a rotation by any angle up to precision $ε$ in T-depth exactly 1 using a universal $\\lceil\\log_2(8/ε)\\rceil$-qubit catalyst state.",
      "url": "http://arxiv.org/abs/2512.24982v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24982v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 636
    },
    {
      "title": "ShowUI-$π$: Flow-based Generative Models as GUI Dexterous Hands",
      "authors": [
        "Siyuan Hu",
        "Kevin Qinghong Lin",
        "Mike Zheng Shou"
      ],
      "abstract": "Building intelligent agents capable of dexterous manipulation is essential for achieving human-like automation in both robotics and digital environments. However, existing GUI agents rely on discrete click predictions (x,y), which prohibits free-form, closed-loop trajectories (e.g. dragging a progress bar) that require continuous, on-the-fly perception and adjustment. In this work, we develop ShowUI-$π$, the first flow-based generative model as GUI dexterous hand, featuring the following designs: (i) Unified Discrete-Continuous Actions, integrating discrete clicks and continuous drags within a shared model, enabling flexible adaptation across diverse interaction modes; (ii) Flow-based Action Generation for drag modeling, which predicts incremental cursor adjustments from continuous visual observations via a lightweight action expert, ensuring smooth and stable trajectories; (iii) Drag Training data and Benchmark, where we manually collect and synthesize 20K drag trajectories across five domains (e.g. PowerPoint, Adobe Premiere Pro), and introduce ScreenDrag, a benchmark with comprehensive online and offline evaluation protocols for assessing GUI agents' drag capabilities. Our experiments show that proprietary GUI agents still struggle on ScreenDrag (e.g. Operator scores 13.27, and the best Gemini-2.5-CUA reaches 22.18). In contrast, ShowUI-$π$ achieves 26.98 with only 450M parameters, underscoring both the difficulty of the task and the effectiveness of our approach. We hope this work advances GUI agents toward human-like dexterous control in digital world. The code is available at https://github.com/showlab/showui-pi.",
      "url": "http://arxiv.org/abs/2512.24965v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24965v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.AI",
        "cs.HC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 637
    },
    {
      "title": "High-performance quantum interconnect between bosonic modules beyond transmission loss constraints",
      "authors": [
        "Hongwei Huang",
        "Jie Zhou",
        "Weizhou Cai",
        "Weiting Wang",
        "Yilong Zhou",
        "Yunlai Zhu",
        "Ziyue Hua",
        "Yifang Xu",
        "Lida Sun",
        "Juan Song"
      ],
      "abstract": "Distributed quantum computing architectures require high-performance quantum interconnects between quantum information processing units, while previous implementations have been fundamentally limited by transmission line losses. Here, we demonstrate a low-loss interconnect between two superconducting modules using an aluminum coaxial cable, achieving a bus mode quality factor of 1.7e6. By employing SNAIL as couplers, we realize inter-modular state transfer in 0.8 μs via a three-wave mixing process. The state transfer fidelity reaches 98.2% for quantum states encoded in the first two energy levels, achieving a Bell state fidelity of 92.5%. Furthermore, we show the capability to transfer high-dimensional states by successfully transmitting binomially encoded logical states. Systematic characterization reveals that performance constraints have shifted from transmission line losses (contributing merely 0.2% infidelity) to module-channel interface effects and local Kerr nonlinearities. Our work advances the realization of quantum interconnects approaching fundamental capacity limits, paving the way for scalable distributed quantum computing and efficient quantum communications.",
      "url": "http://arxiv.org/abs/2512.24926v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24926v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 638
    },
    {
      "title": "Adaptive Resource Orchestration for Distributed Quantum Computing Systems",
      "authors": [
        "Kuan-Cheng Chen",
        "Felix Burt",
        "Nitish K. Panigrahy",
        "Kin K. Leung"
      ],
      "abstract": "Scaling quantum computing beyond a single device requires networking many quantum processing units (QPUs) into a coherent quantum-HPC system. We propose the Modular Entanglement Hub (ModEn-Hub) architecture: a hub-and-spoke photonic interconnect paired with a real-time quantum network orchestrator. ModEn-Hub centralizes entanglement sources and shared quantum memory to deliver on-demand, high-fidelity Bell pairs across heterogeneous QPUs, while the control plane schedules teleportation-based non-local gates, launches parallel entanglement attempts, and maintains a small ebit cache. To quantify benefits, we implement a lightweight, reproducible Monte Carlo study under realistic loss and tight round budgets, comparing a naive sequential baseline to an orchestrated policy with logarithmically scaled parallelism and opportunistic caching. Across 1-128 QPUs and 2,500 trials per point, ModEn-Hub-style orchestration sustains about 90% teleportation success while the baseline degrades toward about 30%, at the cost of higher average entanglement attempts (about 10-12 versus about 3). These results provide clear, high-level evidence that adaptive resource orchestration in the ModEn-Hub enables scalable and efficient quantum-HPC operation on near-term hardware.",
      "url": "http://arxiv.org/abs/2512.24902v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24902v1",
      "citations": null,
      "categories": [
        "quant-ph",
        "cs.DC"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 639
    },
    {
      "title": "Semi-Automated Data Annotation in Multisensor Datasets for Autonomous Vehicle Testing",
      "authors": [
        "Andrii Gamalii",
        "Daniel Górniak",
        "Robert Nowak",
        "Bartłomiej Olber",
        "Krystian Radlak",
        "Jakub Winter"
      ],
      "abstract": "This report presents the design and implementation of a semi-automated data annotation pipeline developed within the DARTS project, whose goal is to create a large-scale, multimodal dataset of driving scenarios recorded in Polish conditions. Manual annotation of such heterogeneous data is both costly and time-consuming. To address this challenge, the proposed solution adopts a human-in-the-loop approach that combines artificial intelligence with human expertise to reduce annotation cost and duration. The system automatically generates initial annotations, enables iterative model retraining, and incorporates data anonymization and domain adaptation techniques. At its core, the tool relies on 3D object detection algorithms to produce preliminary annotations. Overall, the developed tools and methodology result in substantial time savings while ensuring consistent, high-quality annotations across different sensor modalities. The solution directly supports the DARTS project by accelerating the preparation of large annotated dataset in the project's standardized format, strengthening the technological base for autonomous vehicle research in Poland.",
      "url": "http://arxiv.org/abs/2512.24896v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24896v1",
      "citations": null,
      "categories": [
        "cs.AI"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 640
    },
    {
      "title": "Feature Slice Matching for Precise Bug Detection",
      "authors": [
        "Ke Ma",
        "Jianjun Huang",
        "Wei You",
        "Bin Liang",
        "Jingzheng Wu",
        "Yanjun Wu",
        "Yuanjun Gong"
      ],
      "abstract": "Measuring the function similarity to detect bugs is effective, but the statements unrelated to the bugs can impede the performance due to the noise interference. Suppressing the noise interference in existing works does not manage the tough job, i.e., eliminating the noise in the targets. In this paper, we propose MATUS to mitigate the target noise for precise bug detection based on similarity measurement. Feature slices are extracted from both the buggy query and the targets to represent the semantic feature of (potential) bug logics. In particular, MATUS guides the target slicing with the prior knowledge from the buggy code, in an end-to-end way to pinpoint the slicing criterion in the targets. All feature slices are embedded and compared based on the vector similarity. Buggy candidates are audited to confirm unknown bugs in the targets. Experiments show that MATUS holds advantages in bug detection for real-world projects with acceptable efficiency. In total, MATUS has spotted 31 unknown bugs in the Linux kernel. All of them have been confirmed by the kernel developers, and 11 have been assigned CVEs.",
      "url": "http://arxiv.org/abs/2512.24858v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24858v1",
      "citations": null,
      "categories": [
        "cs.SE"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 641
    },
    {
      "title": "friends.test: rank-based method for feature selection in interaction matrices",
      "authors": [
        "Alexandra Suvorikova",
        "Alexey Kroshnin",
        "Dmirijs Lvovs",
        "Vera Mukhina",
        "Andrey Mironov",
        "Elana J. Fertig",
        "Ludmila Danilova",
        "Alexander Favorov"
      ],
      "abstract": "The analysis of the interaction matrix between two distinct sets is essential across diverse fields, from pharmacovigilance to transcriptomics. Not all interactions are equally informative: a marker gene associated with a few specific biological processes is more informative than a highly expressed non-specific gene associated with most observed processes. Identifying these interactions is challenging due to background connections. Furthermore, data heterogeneity across sources precludes universal identification criteria.\n  To address this challenge, we introduce \\textsf{friends.test}, a method for identifying specificity by detecting structural breaks in entity interactions. Rank-based representation of the interaction matrix ensures invariance to heterogeneous data and allows for integrating data from diverse sources. To automatically locate the boundary between specific interactions and background activity, we employ model fitting. We demonstrate the applicability of \\textsf{friends.test} on the GSE112026 -- transnational data from head and neck cancer. A computationally efficient \\textsf{R} implementation is available at https://github.com/favorov/friends.test.",
      "url": "http://arxiv.org/abs/2512.24843v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24843v1",
      "citations": null,
      "categories": [
        "q-bio.QM"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 642
    },
    {
      "title": "Scalable Stellar Parameter Inference Using Python-based LASP: From CPU Optimization to GPU Acceleration",
      "authors": [
        "Jun-Chao Liang",
        "Yin-Bi Li",
        "A-Li Luo",
        "Fang Zuo",
        "Bing Du",
        "Shuo Li",
        "Xiao-Xiao Ma",
        "Shu-Guo Ma",
        "Hai-Ling Lu",
        "Ke-Fei Wu"
      ],
      "abstract": "To enhance the efficiency, scalability, and cross-survey applicability of stellar parameter inference in large spectroscopic datasets, we present a modular, parallelized Python framework with automated error estimation, built on the LAMOST Atmospheric Parameter Pipeline (LASP) originally implemented in IDL. Rather than a direct code translation, this framework refactors LASP with two complementary modules: LASP-CurveFit, a new implementation of the LASP fitting procedure that runs on a CPU, preserving legacy logic while improving data I/O and multithreaded execution efficiency; and LASP-Adam-GPU, a GPU-accelerated method that introduces grouped optimization by constructing a joint residual function over multiple observed and model spectra, enabling high-throughput parameter inference across tens of millions of spectra. Applied to 10 million LAMOST spectra, the framework reduces runtime from 84 to 48 hr on the same CPU platform and to 7 hr on an NVIDIA A100 GPU, while producing results consistent with those from the original pipeline. The inferred errors agree well with the parameter variations from repeat observations of the same target (excluding radial velocities), while the official empirical errors used in LASP are more conservative. When applied to DESI DR1, our effective temperatures and surface gravities agree better with APOGEE than those from the DESI pipeline, particularly for cool giants, while the latter performs slightly better in radial velocity and metallicity. These results suggest that the framework delivers reliable accuracy, efficiency, and transferability, offering a practical approach to parameter inference in large spectroscopic surveys. The code and DESI-based catalog are available via \\dataset[DOI: 10.12149/101679]{https://doi.org/10.12149/101679} and \\dataset[DOI: 10.12149/101675]{https://doi.org/10.12149/101675}, respectively.",
      "url": "http://arxiv.org/abs/2512.24840v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": "10.3847/1538-4357/ae1446",
      "pdf_url": "https://arxiv.org/pdf/2512.24840v1",
      "citations": null,
      "categories": [
        "astro-ph.GA"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 643
    },
    {
      "title": "Non-Abelian Geometric Phases in Triangular Structures And Universal SU(2) Control in Shape Space",
      "authors": [
        "J. Dai",
        "A. Molochkov",
        "A. J. Niemi",
        "J. Westerholm"
      ],
      "abstract": "We construct holonomic quantum gates for qubits that are encoded in the near-degenerate vibrational $E$-doublet of a deformable three-body system. Using Kendall's shape theory, we derive the Wilczek--Zee connection governing adiabatic transport within the $E$-manifold. We show that its restricted holonomy group is $\\mathrm{SU}(2)$, implying universal single-qubit control by closed loops in shape space. We provide explicit loops implementing a $π/2$ phase gate and a Hadamard-type gate. For two-qubit operations, we outline how linked holonomic cycles in arrays generate a controlled Chern--Simons phase, enabling an entangling controlled-$X$ (CNOT) gate. We present a Ramsey/echo interferometric protocol that measures the Wilson loop trace of the Wilczek--Zee connection for a control cycle, providing a gauge-invariant signature of the non-Abelian holonomy. As a physically realizable demonstrator, we propose bond-length modulations of a Cs($6s$)--Cs($6s$)--Cs($nd_{3/2}$)\n  Rydberg trimer in optical tweezers and specify operating conditions that suppress leakage out of the $E$-manifold.",
      "url": "http://arxiv.org/abs/2512.24798v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24798v1",
      "citations": null,
      "categories": [
        "quant-ph",
        "cond-mat.other"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 644
    },
    {
      "title": "Nonlinear Noise2Noise for Efficient Monte Carlo Denoiser Training",
      "authors": [
        "Andrew Tinits",
        "Stephen Mann"
      ],
      "abstract": "The Noise2Noise method allows for training machine learning-based denoisers with pairs of input and target images where both the input and target can be noisy. This removes the need for training with clean target images, which can be difficult to obtain. However, Noise2Noise training has a major limitation: nonlinear functions applied to the noisy targets will skew the results. This bias occurs because the nonlinearity makes the expected value of the noisy targets different from the clean target image. Since nonlinear functions are common in image processing, avoiding them limits the types of preprocessing that can be performed on the noisy targets. Our main insight is that certain nonlinear functions can be applied to the noisy targets without adding significant bias to the results. We develop a theoretical framework for analyzing the effects of these nonlinearities, and describe a class of nonlinear functions with minimal bias.\n  We demonstrate our method on the denoising of high dynamic range (HDR) images produced by Monte Carlo rendering. Noise2Noise training can have trouble with HDR images, where the training process is overwhelmed by outliers and performs poorly. We consider a commonly used method of addressing these training issues: applying a nonlinear tone mapping function to the model output and target images to reduce their dynamic range. This method was previously thought to be incompatible with Noise2Noise training because of the nonlinearities involved. We show that certain combinations of loss functions and tone mapping functions can reduce the effect of outliers while introducing minimal bias. We apply our method to an existing machine learning-based Monte Carlo denoiser, where the original implementation was trained with high-sample count reference images. Our results approach those of the original implementation, but are produced using only noisy training data.",
      "url": "http://arxiv.org/abs/2512.24794v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": "10.1145/3757377.3763931",
      "pdf_url": "https://arxiv.org/pdf/2512.24794v1",
      "citations": null,
      "categories": [
        "cs.CV",
        "cs.GR",
        "cs.LG"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 645
    },
    {
      "title": "Digitalizing Over-the-Air Computation via The Novel Complement Coded Modulation",
      "authors": [
        "Zhixu Wang",
        "Jiacheng Yao",
        "Wei Xu",
        "Wei Shi",
        "Kaibin Huang"
      ],
      "abstract": "To overcome inherent limitations of analog signals in over-the-air computation (AirComp), this letter proposes a two's complement-based coding scheme for the AirComp implementation with compatible digital modulations. Specifically, quantized discrete values are encoded into binary sequences using the two's complement and transmitted over multiple subcarriers. At the receiver, we design a decoder that constructs a functional mapping between the superimposed digital modulation signals and the target of computational results, theoretically ensuring asymptotic error free computation with the minimal codeword length. To further mitigate the adverse effects of channel fading, we adopt a truncated inversion strategy for pre-processing. Benefiting from the unified symbol distribution after the proposed encoding, we derive the optimal linear minimum mean squared error (LMMSE) detector in closed form and propose a low complexity algorithm seeking for the optimal truncation selection. Furthermore, the inherent importance differences among the coded outputs motivate an uneven power allocation strategy across subcarriers to improve computational accuracy. Numerical results validate the superiority of the proposed scheme over existing digital AirComp approaches, especially at low signal to-noise ratio (SNR) regimes.",
      "url": "http://arxiv.org/abs/2512.24788v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24788v1",
      "citations": null,
      "categories": [
        "eess.SP"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 646
    },
    {
      "title": "Runaway electron avalanche and macroscopic beam formation: simulations of the DTT full power scenario",
      "authors": [
        "E. Emanuelli",
        "F. Vannini",
        "M. Hoelzl",
        "E. Nardon",
        "V. Bandaru",
        "N. Schwarz",
        "D. Bonfiglio",
        "G. Ramogida",
        "F. Subba",
        "JOREK Team"
      ],
      "abstract": "The transition of the Divertor Tokamak Test (DTT) facility from its initial commissioning phase (Day-0, plasma current $I_{p}=2$ MA) to the full power scenario ($I_{p}=5.5$ MA) introduces a critical shift in the dynamics of runaway electrons (REs) generation. While previous predictive studies of the low-current scenario indicated a robust safety margin against RE beam formation, this work reveals that the exponential scaling of the RE avalanche gain with plasma current severely narrows the safe operational window in the full power scenario. Using the non-linear magnetohydrodynamic code JOREK, we perform comprehensive 2D simulations of the current quench (CQ) phase of several disruption scenarios, systematically scanning initial RE seed currents and injected impurity levels. The results demonstrate that in the full power scenario, the avalanche multiplication factor is sufficiently high ($G_\\text{av} \\approx 1.3 \\cdot 10^5$) to convert a mere 5.5 A seed current into macroscopic RE beams of $\\approx 0.7$ MA when large amounts of impurities are present. For even higher RE seeds, the RE current can peak at $ \\approx 3.2$ MA, constituting up to $\\approx$ 80% of the total plasma current during the CQ. These findings suggest that, unlike the Day-0 phase, the disruption mitigation strategy for the full power scenario involves a careful balance between thermal load mitigation and RE avoidance, necessitating a well-chosen quantity of injected impurities. This work provides the baseline needed for future estimations of RE loads on the plasma-facing components of DTT, which will be essential for designing and positioning mitigation components like sacrificial limiters.",
      "url": "http://arxiv.org/abs/2512.24760v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24760v1",
      "citations": null,
      "categories": [
        "physics.plasm-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 647
    },
    {
      "title": "Easier randomizing gates provide more accurate fidelity estimation",
      "authors": [
        "Debankan Sannamoth",
        "Kristine Boone",
        "Arnaud Carignan-Dugas",
        "Akel Hashim",
        "Irfan Siddiqi",
        "Karl Mayer",
        "Joseph Emerson"
      ],
      "abstract": "Accurate benchmarking of quantum gates is crucial for understanding and enhancing the performance of quantum hardware. A standard method for this is interleaved benchmarking, a technique which estimates the error on an interleaved target gate by comparing cumulative error rates of randomized sequences implemented with the interleaved gate and without it. In this work, we show both numerically and experimentally that the standard approach of interleaved randomized benchmarking (IRB), which uses the multi-qubit Clifford group for randomization, can produce highly inaccurate and even physically impossible estimates for the error on the interleaved gate in the presence of coherent errors. Fortunately we also show that interleaved benchmarking performed with cycle benchmarking, which randomizes with single qubit Pauli gates, provides dramatically reduced systematic uncertainty relative to standard IRB, and further provides as host of additional benefits including data reusability. We support our conclusions with a theoretical framework for bounding systematic errors, extensive numerical results comparing a range of interleaved protocols under fixed resource costs, and experimental demonstrations on three quantum computing platforms.",
      "url": "http://arxiv.org/abs/2512.24744v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24744v1",
      "citations": null,
      "categories": [
        "quant-ph"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 648
    },
    {
      "title": "S-Duality for Non-Abelian Monopoles",
      "authors": [
        "Shan Hu"
      ],
      "abstract": "In $\\mathcal{N}=4$ super-Yang-Mills theory with gauge group $G$ spontaneously broken to a subgroup $H$, S-duality requires that the BPS monopole spectrum organizes into the same representation as W-bosons in the dual theory, where $G^{\\vee}$ is broken to $H^{\\vee}$. The expectation has been extensively verified in the maximally broken phase $G\\to U(1)^r$. Here we address the non-Abelian regime in which $H$ contains a semisimple factor $H^{s}$. Using the stratified description of monopole moduli space, we give a general proof of this matching for any simple gauge group $G$. Each BPS monopole state is naturally labeled by a weight of the relevant $W$-boson representation of $(H^{\\vee})^{s}$. We construct non-Abelian magnetic gauge transformation operators implementing the $(H^{\\vee})^{s}$-action on the monopole Hilbert space, which commute with the electric $H^{s}$-transformations and thereby realize the $H^{s}\\times (H^{\\vee})^{s}$ symmetry at the level of monopole quantum mechanics.",
      "url": "http://arxiv.org/abs/2512.24743v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24743v1",
      "citations": null,
      "categories": [
        "hep-th"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 649
    },
    {
      "title": "Splatwizard: A Benchmark Toolkit for 3D Gaussian Splatting Compression",
      "authors": [
        "Xiang Liu",
        "Yimin Zhou",
        "Jinxiang Wang",
        "Yujun Huang",
        "Shuzhao Xie",
        "Shiyu Qin",
        "Mingyao Hong",
        "Jiawei Li",
        "Yaowei Wang",
        "Zhi Wang"
      ],
      "abstract": "The recent advent of 3D Gaussian Splatting (3DGS) has marked a significant breakthrough in real-time novel view synthesis. However, the rapid proliferation of 3DGS-based algorithms has created a pressing need for standardized and comprehensive evaluation tools, especially for compression task. Existing benchmarks often lack the specific metrics necessary to holistically assess the unique characteristics of different methods, such as rendering speed, rate distortion trade-offs memory efficiency, and geometric accuracy. To address this gap, we introduce Splatwizard, a unified benchmark toolkit designed specifically for benchmarking 3DGS compression models. Splatwizard provides an easy-to-use framework to implement new 3DGS compression model and utilize state-of-the-art techniques proposed by previous work. Besides, an integrated pipeline that automates the calculation of key performance indicators, including image-based quality metrics, chamfer distance of reconstruct mesh, rendering frame rates, and computational resource consumption is included in the framework as well. Code is available at https://github.com/splatwizard/splatwizard",
      "url": "http://arxiv.org/abs/2512.24742v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24742v1",
      "citations": null,
      "categories": [
        "cs.CV"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 650
    },
    {
      "title": "Model-independent search of gravitational wave echoes in LVK data",
      "authors": [
        "Di Wu",
        "Xi-Li Zhang",
        "Qing-Guo Huang",
        "Jing Ren"
      ],
      "abstract": "Gravitational wave echoes offer a unique probe of the near-horizon structure of astrophysical black holes, beyond the standard ''black hole spectroscopy''. Theoretical waveform predictions, however, remain uncertain, motivating robust searches that avoid specific echo modeling. We present a model-independent search framework targeting long-lived quasinormal modes (QNMs) expected from strong interior reflection. By employing a generalized phase-marginalized likelihood that coherently combines data for each QNM across a detector network, our method enhances sensitivity to the signals. To handle real detector noise, we implement an optimized notching procedure to suppress instrumental spectral lines and refine the Bayesian parameter settings. We validate the performance of this framework using injection studies on O1 background data, demonstrating reliable signal recovery in realistic noise conditions. We then apply this method to three binary black hole merger events with high ringdown signal-to-noise ratios (SNR) from observing runs O1 to O4: GW150914, GW231226, and the recently detected GW250114. No statistically significant evidence for postmerger echoes is found. Consequently, we derive 90% upper limits on the network SNR and the average amplitude of the long-lived QNMs, setting the first model-independent constraints on late-time echo signatures from LVK data.",
      "url": "http://arxiv.org/abs/2512.24730v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24730v1",
      "citations": null,
      "categories": [
        "gr-qc",
        "astro-ph.HE"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 651
    },
    {
      "title": "Phase transitions in time complexity of Brownian circuits",
      "authors": [
        "Kota Okajima",
        "Koji Hukushima"
      ],
      "abstract": "Brownian circuits implement computation through stochastic transitions driven by thermal fluctuations. While the energetic costs of such fluctuation-driven computation have been extensively studied within stochastic thermodynamics, much less is known about its computational complexity, in particular how computation time scales with circuit size. Here, the computation time of explicitly designed Brownian circuits is investigated numerically via the first-passage time to a completed state. For arithmetic circuits such as adders, varying the forward transition rate induces a sharp change in the scaling behavior of the mean computation time, from linear to exponential in circuit size. This change can be interpreted as an easy-hard transition in computational time complexity. The transition suggests that, for meaningful computational tasks, achieving efficient polynomial-time computation generically requires a finite forward bias, corresponding to a nonzero energy input. As a counterexample, it is shown that arbitrary logical operations can be reduced to an effectively one-dimensional stochastic process, for which the zero-bias limit lines within the computationally efficient (easy) regime. However, realizing such a one-dimensional normal form unavoidably leads to an exponential increase in circuit size. These results reveal a fundamental trade-off between computation time, circuit size, and energy input in Brownian circuits, and demonstrate that phase transitions in time complexity provide a natural framework for characterizing the cost of fluctuation-driven computation.",
      "url": "http://arxiv.org/abs/2512.24728v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24728v1",
      "citations": null,
      "categories": [
        "cond-mat.stat-mech"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 652
    },
    {
      "title": "Equivalence of Personalized PageRank and Successor Representations",
      "authors": [
        "Beren Millidge"
      ],
      "abstract": "The hippocampus appears to implement two core but highly distinct functions in the brain: long term memory retrieval and planning and spatial navigation. Naively, these functions appear very different algorithmically. In this short note, we demonstrate that two powerful algorithms that have each independently been proposed to underlie the hippocampal operation for each function -- personalized page-rank for memory retrieval, and successor representations for planning and navigation, are in fact isomorphic and utilize the same underlying representation -- the stationary distribution of a random walk on a graph. We hypothesize that the core computational function of the hippocampus is to compute this representation on arbitrary input graphs.",
      "url": "http://arxiv.org/abs/2512.24722v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24722v1",
      "citations": null,
      "categories": [
        "cs.NE"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 653
    },
    {
      "title": "Primordial black hole dark matter from ultra-slow-roll inflation in Horndeski gravity",
      "authors": [
        "Despina Totolou",
        "Theodoros Papanikolaou",
        "Emmanuel N. Saridakis"
      ],
      "abstract": "Primordial black holes (PBHs) provide a well-motivated non-particle candidate for dark matter, requiring an enhancement of curvature perturbations on small inflationary scales consistent with observational constraints. In this work we study PBH production within Horndeski gravity, accounting for compatibility with the GW170817 constraint on the gravitational-wave speed and imposing a constant coupling to the Ricci scalar. Under these conditions, and assuming an inflaton field characterised by a canonical kinetic term and a smooth potential, the inflationary dynamics is controlled by the cubic Horndeski interaction. We show that a suitable kinetic dependence of the latter enhances the effective friction acting on the inflaton, inducing a transient ultra-slow-roll phase embedded in an otherwise standard slow-roll evolution. Interestingly, this mechanism amplifies the curvature power spectrum on small scales without introducing any feature in the potential. For representative parameter choices we find that pronounced peaks in the scalar power spectrum are generated, leading to the formation of asteroid-mass PBHs with masses of order $\\mathcal{O}(10^{-16})\\,M_\\odot$, which can account for a substantial fraction of the dark matter abundance, reaching $f_{\\rm PBH}\\simeq 0.9$, while satisfying current observational constraints. The resulting sharp features in the scalar power spectrum also imply potentially observable scalar-induced gravitational-wave signatures.",
      "url": "http://arxiv.org/abs/2512.25044v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25044v1",
      "citations": null,
      "categories": [
        "gr-qc",
        "astro-ph.CO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 654
    },
    {
      "title": "Towards precision cosmology with Voids x CMB correlations (I): Roman-Agora mock catalogs and pipeline validation",
      "authors": [
        "Mar Pérez Sar",
        "Carlos Hernández Monteagudo",
        "András Kovács",
        "Alice Pisani"
      ],
      "abstract": "We construct and validate a set of multi-purpose mock galaxy catalogs designed to capture, to different degrees of accuracy, the main characteristics of the Nancy Grace Roman Space Telescope survey. These catalogs provide a foundation for void statistics and various CMB cross-correlation analyses. Our approach differs from traditional halo occupation or abundance matching methods by directly translating a reference mock catalog -- containing basic properties of the host halos -- into a new simulation (in our case Agora). This technique, which we call analog matching, assigns a halo counterpart in the new simulation to each reference galaxy through a nearest-neighbor search in a multi-dimensional parameter space. This space can include halo mass, environmental measures and other galaxy-specific attributes. By varying the composition of this parameter vector, we can generate catalogs of differing complexity and conduct systematic tests to examine the influence of modelling choices on LSS statistics. We find that analog matching based on halo mass alone, or halo mass and galaxy-type indicators, successfully reproduces the expected Roman emission-line galaxy statistics. We also show that reproducing two-dimensional galaxy clustering does not guarantee consistent void properties. Our results highlight the importance of matching void statistics for improved mock accuracy, and demonstrate that measuring voids provides independent and sensitive constraints on galaxy-halo connections beyond the matter power spectrum. An important by-product of our setup is that it is fully general and can be applied to any combination of simulation and reference catalog, provided that the desired parameter space for both is specified. The resulting Roman-Agora mock catalogs offer a versatile resource for LSS x CMB studies and a benchmark for assessing the impact of mock accuracy on cosmological observables.",
      "url": "http://arxiv.org/abs/2512.25040v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.25040v1",
      "citations": null,
      "categories": [
        "astro-ph.CO"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 655
    },
    {
      "title": "Fair Committee Selection under Ordinal Preferences and Limited Cardinal Information",
      "authors": [
        "Ameet Gadekar",
        "Aristides Gionis",
        "Suhas Thejaswi",
        "Sijing Tu"
      ],
      "abstract": "We study the problem of fair $k$-committee selection under an egalitarian objective. Given $n$ agents partitioned into $m$ groups (\\eg, demographic quotas), the goal is to aggregate their preferences to form a committee of size $k$ that guarantees minimum representation from each group while minimizing the maximum \\emph{cost} incurred by any agent. We model this setting as the ordinal fair $k$-center problem, where agents are embedded in an unknown metric space, and each agent reports a complete preference ranking (i.e., ordinal information) over all agents, consistent with the underlying distance metric (i.e., cardinal information). The cost incurred by an agent with respect to a committee is defined as its distance to the closest committee member. The quality of an algorithm is evaluated using the notion of distortion, which measures the worst-case ratio between the cost of the committee produced by the algorithm and the cost of an optimal committee, when given complete access to the underlying metric space.\n  When cardinal information is not available, no constant distortion is possible for the ordinal $k$-center problem, even without fairness constraints, when $k\\geq 3$ [Burkhardt et.al., AAAI'24]. To overcome this hardness, we allow limited access to cardinal information by querying the metric space. In this setting, our main contribution is a factor-$5$ distortion algorithm that requires only $O(k \\log^2 k)$ queries. Along the way, we present an improved factor-$3$ distortion algorithm using $O(k^2)$ queries.",
      "url": "http://arxiv.org/abs/2512.24934v1",
      "year": 2025,
      "venue": "arXiv",
      "source": "arxiv",
      "doi": null,
      "pdf_url": "https://arxiv.org/pdf/2512.24934v1",
      "citations": null,
      "categories": [
        "cs.DS"
      ],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 656
    },
    {
      "title": "Single Phase Immersion Cooling for Hyper Scale Data Centers: Challenges and Opportunities",
      "authors": [
        "D. Agonafer",
        "P. Bansode",
        "S. Saini",
        "J. Gullbrand",
        "Ashish Gupta"
      ],
      "abstract": "\n Rapidly expanding computing, storage, and networking requirements have increased the quantity and energy density of modern data centers. Air-cooled high-performance servers often require low air-supply temperatures as well as high air-flow rates, making air-cooling inefficient above certain thermal design power levels. Single-phase liquid immersion cooling (Sp-LIC) addresses this challenge by providing a much higher thermal mass and a high percentage of heat dissipation owing to the direct contact of dielectric fluids with all powered components in the server. It also improves reliability by shielding the ITE from pollutants and hostile environments, and it lowers OpEx by eliminating fans and computer room air handling units. Unlike direct-to-chip liquid cooling, Sp-LIC does not need a complex liquid distribution manifold design as the dielectric liquid can be in direct contact with all components of a server, making it an ideal choice for hyper-scale, edge, and modular data center applications. There are some unique thermal challenges to implementing Sp-LIC such as determining whether to use natural or forced convection and customizing heat sinks that were designed for air-cooling that need to be optimized for dielectric fluids with higher fin efficiency and component reliability (active and passive), which must be addressed and thoroughly researched. These challenges can be addressed using CFD simulations including “Multi Design Variables” and “Multi-Objective Function” Optimization with “Constraints” to help in the design of appropriate extended surfaces and cooling approaches. Experiments are done to verify the CFD models for inlet temperatures greater than 40°C. Also, there are very limited studies related to the reliability of such cooling technology. The accelerated thermal cycling (ATC) test given by ATC JEDEC is relevant just for air cooling but there is no such standard for immersion cooling. The ASTM benchmark D3455 with some appropriate adjustments was adopted to test the material compatibility because of the air and dielectric fluid differences in the heat capacitance property and corresponding ramp rate during thermal cycling. Material characterization findings such as modulus, CTE, creep, fracture toughness, elastic modulus, stress relaxation, cracking, dislocation nucleation, and the viscoelastic properties of the samples will be compared before and after immersion using a TI-980 Nano-indenter, Dynamic Mechanical Analyzer (DMA), Thermomechanical Analyzer (TMA), and Scanning Electron Microscopy (SEM). In all, a comprehensive guide to Sp-LIC for the thermal management of hyper-scale data centers will be presented.",
      "url": "https://www.semanticscholar.org/paper/38d1cd159f40d43683c2380abae8cd82d15a4465",
      "year": 2023,
      "venue": "ASME 2023 Heat Transfer Summer Conference",
      "source": "semantic_scholar",
      "doi": "10.1115/ht2023-107598",
      "pdf_url": "",
      "citations": 3,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 657
    },
    {
      "title": "Logarithmic Connections on Principal Bundles and Their Applications to Geometric Control Theory",
      "authors": [
        "Álvaro Antón‐Sancho"
      ],
      "abstract": "In this research, we establish a precise correspondence between the theory of logarithmic connections on principal G-bundles over compact Riemann surfaces and the geometric formulation of control systems on curved manifolds, providing a novel differential–geometric framework for analyzing optimal control problems with non-holonomic constraints. By characterizing control systems through the geometric structure of flat connections with logarithmic singularities at marked points, we demonstrate that optimal trajectories correspond precisely to horizontal lifts with respect to the connection. These horizontal lifts project onto geodesics on the punctured surface, which is equipped with a Riemannian metric uniquely determined by the monodromy representation around the singularities. The main geometric result proves that the isomonodromic deformation condition translates into a compatibility condition for the control system. This condition preserves the conjugacy classes of monodromy transformations under variations of the marked points, and ensures the existence and uniqueness of optimal trajectories satisfying prescribed boundary conditions. Furthermore, we analyze systems with non-holonomic constraints by relating the constraint distribution to the kernel of the connection form, showing how the degree of non-holonomy can be measured through the failure of integrability of the associated horizontal distribution on the principal bundle. As an application, we provide computational implementations for SL(2,C) connections over hyperbolic Riemann surfaces with genus g≥2, explicitly constructing the monodromy-induced metric via the Poincaré uniformization theorem and deriving closed-form expressions for optimal control strategies that exhibit robust performance characteristics under perturbations of initial conditions and system parameters.",
      "url": "https://www.semanticscholar.org/paper/cd7b2a29f28bf5b52ca25ed9e37466cd5a44fad2",
      "year": 2025,
      "venue": "Axioms",
      "source": "semantic_scholar",
      "doi": "10.3390/axioms15010010",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 658
    },
    {
      "title": "AutoMap: Automatic Mapping of Neural Networks to Deep Learning Accelerators for Edge Devices",
      "authors": [
        "Yanhong Wang",
        "Zihao Zhao",
        "Xu Jin",
        "Haotian Zheng",
        "Maohua Nie",
        "Qiaosha Zou",
        "Chenlu Shi"
      ],
      "abstract": "Emerging deep neural networks (DNNs) have been emerging in applications (object detection, automatic speech recognition, etc.) deployed on edge devices. To improve the energy efficiency of edge devices, domain-specific deep learning accelerators (DLAs) are designed with limited on-chip resources. The manifold DLA designs and evolving DNN topologies bring challenges for applications mapping and scheduling on hardware resources. In this article, we propose an automatic DNN mapping framework named AutoMap, given the hardware backend information. First, a computational graph representation called extended directed weighted graph (EDWG) is proposed, which realizes unified expression for both spatial and temporal network interlayer connections. Second, an associated partitioner is implemented for splitting an EDWG into subEDWGs, which incorporates the on-chip memory constraint and facilitates weight data reuse on chip. Finally, a dynamic memory allocation strategy is utilized to alleviate the feature storing burden introduced by the multivarious network sizes and connections. Compared to the baseline mapping methods, experimental results show that our proposed automatic mapping framework can help to speedup the execution of several DNNs on state-of-the-art DLAs, ranging from $1.27\\times $ to $3.45\\times $ . The utilization of the PE array can increase from 20% to 64%.",
      "url": "https://www.semanticscholar.org/paper/b7bf209183e78a4e8253f79b95847dece9925401",
      "year": 2023,
      "venue": "IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems",
      "source": "semantic_scholar",
      "doi": "10.1109/TCAD.2022.3232070",
      "pdf_url": "",
      "citations": 4,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 659
    },
    {
      "title": "Time Makes Space: Emergence of Place Fields in Networks Encoding Temporally Continuous Sensory Experiences",
      "authors": [
        "Zhaoze Wang",
        "Ronald W. Di Tullio",
        "Spencer Rooke",
        "Vijay Balasubramanian"
      ],
      "abstract": "The vertebrate hippocampus is believed to use recurrent connectivity in area CA3 to support episodic memory recall from partial cues. This brain area also contains place cells, whose location-selective firing fields implement maps supporting spatial memory. Here we show that place cells emerge in networks trained to remember temporally continuous sensory episodes. We model CA3 as a recurrent autoencoder that recalls and reconstructs sensory experiences from noisy and partially occluded observations by agents traversing simulated arenas. The agents move in realistic trajectories modeled from rodents and environments are modeled as continuously varying, high-dimensional, sensory experience maps (spatially smoothed Gaussian random fields). Training our autoencoder to accurately pattern-complete and reconstruct sensory experiences with a constraint on total activity causes spatially localized firing fields, i.e., place cells, to emerge in the encoding layer. The emergent place fields reproduce key aspects of hippocampal phenomenology: a) remapping (maintenance of and reversion to distinct learned maps in different environments), implemented via repositioning of experience manifolds in the network’s hidden layer, b) orthogonality of spatial representations in different arenas, c) robust place field emergence in differently shaped rooms, with single units showing multiple place fields in large or complex spaces, and d) slow representational drift of place fields. We argue that these results arise because continuous traversal of space makes sensory experience temporally continuous. We make testable predictions: a) rapidly changing sensory context will disrupt place fields, b) place fields will form even if recurrent connections are blocked, but reversion to previously learned representations upon remapping will be abolished, c) the dimension of temporally smooth experience sets the dimensionality of place fields, including during virtual navigation of abstract spaces. Code for our experiments is available at1.",
      "url": "https://www.semanticscholar.org/paper/2c63b474de44a4982b68b8d321427ebef6f85574",
      "year": 2024,
      "venue": "Neural Information Processing Systems",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2408.05798",
      "pdf_url": "",
      "citations": 9,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 660
    },
    {
      "title": "Riemannian Bilevel Optimization",
      "authors": [
        "Sanchayan Dutta",
        "Xiang Cheng",
        "S. Sra"
      ],
      "abstract": "We develop new algorithms for Riemannian bilevel optimization. We focus in particular on batch and stochastic gradient-based methods, with the explicit goal of avoiding second-order information such as Riemannian hyper-gradients. We propose and analyze $\\mathrm{RF^2SA}$, a method that leverages first-order gradient information to navigate the complex geometry of Riemannian manifolds efficiently. Notably, $\\mathrm{RF^2SA}$ is a single-loop algorithm, and thus easier to implement and use. Under various setups, including stochastic optimization, we provide explicit convergence rates for reaching $\\epsilon$-stationary points. We also address the challenge of optimizing over Riemannian manifolds with constraints by adjusting the multiplier in the Lagrangian, ensuring convergence to the desired solution without requiring access to second-order derivatives.",
      "url": "https://www.semanticscholar.org/paper/45a2b6a8621acb3a2f1446c1cd8e56180cca1e44",
      "year": 2024,
      "venue": "arXiv.org",
      "source": "semantic_scholar",
      "doi": "10.48550/arXiv.2405.15816",
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 661
    },
    {
      "title": "A note on the Hamiltonian structure of transgression forms",
      "authors": [
        "P. Pais",
        "Patricio Salgado-Rebolledo",
        "Aldo Vera"
      ],
      "abstract": "By incorporating two gauge connections, transgression forms provide a generalization of Chern-Simons actions that are genuinely gauge-invariant on bounded manifolds. In this work, we show that, when defined on a manifold with a boundary, the Hamiltonian formulation of a transgression field theory can be consistently carried out without the need to implement regularizing boundary terms at the level of first-class constraints. By considering boundary variations of the relevant functionals in the Poisson brackets, the surface integral in the very definition of a transgression action can be translated into boundary contributions in the generators of gauge transformations and diffeomorphisms. This prescription systematically leads to the corresponding surface charges of the theory, reducing to the general expression for conserved charges in (higher-dimensional) Chern-Simons theories when one of the gauge connections in the transgression form is set to zero.",
      "url": "https://www.semanticscholar.org/paper/b10b717502df0fae1fc8bc71e4a13c5398944109",
      "year": 2023,
      "venue": "Journal of High Energy Physics",
      "source": "semantic_scholar",
      "doi": "10.1007/JHEP12(2023)190",
      "pdf_url": "https://link.springer.com/content/pdf/10.1007/JHEP12(2023)190.pdf",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 662
    },
    {
      "title": "A Novel Dynamic Hybrid Beamforming Design for ELAA Systems",
      "authors": [
        "Meng-Ting Liu",
        "Ming Li",
        "Rang Liu",
        "Qian Liu"
      ],
      "abstract": "Extremely large-scale antenna array (ELAA) is deemed as one of several key candidate technologies for the sixth generation (6G) mobile networks. Nevertheless, the near-field effect poses a significant challenge for ELAA systems as a result of employing a substantial quantity of antennas for the transmission of high-frequency signals. Furthermore, the practical implementation of ELAA by employing a hybrid beamforming framework boosts the impact of the near-field effect on the system performance. In order to effectively address this severe near-field effect, we propose a novel dynamic hybrid beamforming architecture, in which each antenna is either adaptively connected to one radio frequency (RF) chain for signal transmission, or deactivated for power saving. The dynamic hybrid beamforming design algorithm is developed to maximize the achievable sum-rate under the constraints of the constant modulus of phase shifters, non-overlapping dynamic connection network, and the total transmit power. To address the resulting complicated nonconvex design problem, we employ the fractional programming (FP) method to transform the objective function into a more tractable form and exploit the manifold-based algorithm to tackle the nonconvex constraint. The simulation results demonstrate the advantages of the proposed dynamic hybrid beamforming and the effectiveness of the FP-manifold-based algorithm in ELAA near-field communication systems.",
      "url": "https://www.semanticscholar.org/paper/185bbc06d2122e14ca4d5d72908ddf708ddd4e4d",
      "year": 2024,
      "venue": "ICC 2024 - IEEE International Conference on Communications",
      "source": "semantic_scholar",
      "doi": "10.1109/ICC51166.2024.10622286",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 663
    },
    {
      "title": "GNPHE / 03-04 hep-th / 0303198 M-theory on G 2 manifolds and the method of ( p , q ) brane webs",
      "authors": [
        "A. Belhaj"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/dde5ad445fcd6894ccddaa7c3ebe9f02420939f9",
      "year": 2022,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 664
    },
    {
      "title": "Warmer for Less: A Cost-Efficient Strategy for Cold-Start Recommendations at Pinterest",
      "authors": [
        "Saeed Ebrahimi",
        "Weijie Jiang",
        "Jaewon Yang",
        "Olafur Gudmundsson",
        "Yucheng Tu",
        "Huizhong Duan"
      ],
      "abstract": "Pinterest is a leading visual discovery platform where recommender systems (RecSys) are key to delivering relevant, engaging, and fresh content to our users. In this paper, we study the problem of improving RecSys model predictions for cold-start (CS) items, which appear infrequently in the training data. Although this problem is well-studied in academia, few studies have addressed its root causes effectively at the scale of a platform like Pinterest. By investigating live traffic data, we identified several challenges of the CS problem and developed a corresponding solution for each: First, industrial-scale RecSys models must operate under tight computational constraints. Since CS items are a minority, any related improvements must be highly cost-efficient. To address this, our solutions were designed to be lightweight, collectively increasing the total parameters by only 5%. Second, CS items are represented only by non-historical (e.g., content or attribute) features, which models often treat as less important. To elevate their significance, we introduce a residual connection for the non-historical features. Third, CS items tend to receive lower prediction scores compared to non-CS items, reducing their likelihood of being surfaced. We mitigate this by incorporating a score regularization term into the model. Fourth, the labels associated with CS items are sparse, making it difficult for the model to learn from them. We apply the manifold mixup technique to address this data sparsity. Implemented together, our methods increased fresh content engagement at Pinterest by 10% without negatively impacting overall engagement and cost, and have been deployed to serve over 570 million users on Pinterest.",
      "url": "https://www.semanticscholar.org/paper/a2dcad40027095c84a3d570f9fe6ccbbcdb6af86",
      "year": 2025,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 665
    },
    {
      "title": "Digital Innovation of Well Construction Process in Ecuador Through Rig Automation",
      "authors": [
        "Karen Peña",
        "Kevin Etcheverry",
        "Hugo Quevedo",
        "Esteban Rojas",
        "R. Correa",
        "José Castellanos",
        "Cory Labatte",
        "Leonardo Márquez",
        "W. Sepúlveda",
        "Brennan Goodkey"
      ],
      "abstract": "\n Artificial intelligence-based (AI) digital drilling technology was implemented in two mature fields of Ecuador, which represent 33% of the country's oil production and where it is essential to maximize the return on investment. The drilling campaign's main strategy included the deployment of a novel automation solution on two rigs, resulting in the optimization of the well construction process.\n In this paper we present the results of implementing a rig automation solution that we applied to 20 wells in 2022. The study primarily focuses on analyzing the results achieved during this period, with the aim of evaluating the solution's effectiveness, performance, and its impact on drilling operations.\n The rig automation system uses a goal-based approach to handle constantly changing drilling conditions like drilling dysfunctions and formation changes and repetitive tasks such as downlinking and pre- and post-connection procedures, without having constantly rely on human intervention. The system sends command packages to the rig control system to direct the surface equipment including the drawworks, automated driller, topdrive, and mud pumps to operate within the constraints of standard operating procedures that are incorporated into the system configuration in the form of procedures, mitigation strategies, and operational limits. It uses powerful data analysis and learning systems to assist and enhance every task from optimizing the rate of penetration (ROP) to drilling a stand autonomously.\n Following the deployment of the rig automation solution on both rigs, the team entered into a hyper care phase, which consisted of monitoring the system and training the drillers at the rig-site over a period of at least 3 months, Once concluded, the team spent the next year leveraging the system and routinely reviewing and optimizing the procedures and configuration parameters to maximize value added. Overall, a total of 20 wells were drilled with the digital solution throughout 2022 including 214,000 feet drilled in automated mode, an on-bottom ROP improvement of 8.4%, with total time savings adding to 23.86 days, and pre- and post-connection time reductions of 54% on average, with latest wells achieving as little as 6 minutes as compared to the initial benchmark of 25 minutes. Additional collaborative features were implemented, such as the remote monitoring and execution of the directional drilling process, where the engineer located in the remote operations center was able to send a downlink request, and once accepted by the driller, the system would dynamically re-plan a new sequence to incorporate the new goal and execute the downlink procedure autonomously while drilling, while the driller monitored other variables to optimize and drive safer operations.\n In the oil and gas industry, automation is mostly associated with the control of heavy machinery that is operated through predefined sequence automation. This goal-based technology provides a significant advantage by minimizing the human interaction needed to deal with the uncertainty of dynamically changing operations and, ultimately, lower drilling times. This indirectly results in a reduction in the carbon emissions footprint, operational risks, and cost, optimizing efficiency and field development.",
      "url": "https://www.semanticscholar.org/paper/f89e3831b1e947962f0f0e1e241737e84975f540",
      "year": 2023,
      "venue": "Day 2 Tue, October 03, 2023",
      "source": "semantic_scholar",
      "doi": "10.2118/216249-ms",
      "pdf_url": "",
      "citations": 1,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 666
    },
    {
      "title": "Combinatorial decompositions for deformed or decorated classes of maps",
      "authors": [
        "V. Nador"
      ],
      "abstract": "The perturbative expansion of tensorial field theories in Feynman graphs can be interpreted as weighted generating series of some piecewise linear varieties. This simple fact establishes a link between two a priori distinct fields: the combinatorics of discrete manifolds on one hand and tensorial field theories on the other hand. In this thesis, we study different aspects revolving around this connection between combinatorics and field theory. First, we consider constellations model, which generalize maps and their algebraic properties. This makes them suited to probe the b-deformation, a deformation of the algebra of symmetric functions. We will study the constraints satisfied by the generating series of cubical b-deformed constellations. Second, we analyze the double scaling limit of particular tensor models of order 3. For tensor of order greater than two, the nature of the 1/N-expansion is qualitatively different from the matrix case of order 2. In particular, only the leading order graphs are fully characterized. Despite this fact, it is possible to identify graphs of subleading orders contributing to the double scaling limit by implementing the scheme decomposition for Feynman graphs of these theories. An analysis of the singularity of the schemes then allows us to give a complete characterization of the graphs contributing to the double scaling limit. Finally, we investigate a particular link between a tensor and a vector field theory which both admit a melonic limit. Namely, we will show that we can obtain the vectorial Amit-Roginski model by considering perturbations around a classical solution of the Boulatov model, a tensorial theory. We give sufficient conditions on the classical solution so that the effective action for the perturbation around this solution takes the form of the Amit-Roginski action.",
      "url": "https://www.semanticscholar.org/paper/e08cf8185b2aedcef2858e849df7737ef10c157e",
      "year": 2023,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 667
    },
    {
      "title": "on Matter under Extreme Conditions in Solar System Giant Planets and Exoplanets, Inverse Problems and Deep Learning",
      "authors": [],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/d6d265d2ea2ff8c251923f3c871a15a51ee53623",
      "year": 2022,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 668
    },
    {
      "title": "Optimal Mass Transport Meets Thermodynamics: On Power and Efficiency of Finite-Time Thermodynamic Engines",
      "authors": [
        "Huidong Chen"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/a3ebed9b1e75a528f8e943275bdb99ffe515ab2c",
      "year": 2022,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 669
    },
    {
      "title": "Projects with allocated PhD studentships Algorithms and Data Analysis",
      "authors": [],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/0f61d7b297f3c6c3652627c8f1d68f5f32afd634",
      "year": 2021,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 670
    },
    {
      "title": "Team Apics Analysis and Problems of Inverse type in Control and Signal processing",
      "authors": [
        "S. Antipolis"
      ],
      "abstract": "",
      "url": "https://www.semanticscholar.org/paper/f6389938d5cc150e359aee87d41e2dd2135a43c8",
      "year": 2021,
      "venue": "",
      "source": "semantic_scholar",
      "doi": null,
      "pdf_url": "",
      "citations": 0,
      "categories": [],
      "id": null,
      "track": null,
      "status": null,
      "keywords": null,
      "tldr": null,
      "primary_area": null,
      "similarity_score": 0.0,
      "novelty_score": 0.0,
      "recency_score": 0.0,
      "relevance_score": 0.0,
      "bm25_score": 0.0,
      "combined_score": 0.0,
      "rank": 671
    }
  ],
  "metadata": {
    "query": "hyper-connections manifold constraint implementation",
    "total_steps": 28,
    "last_updated": "2026-01-03T00:57:34.828924",
    "started_at": "2026-01-03T00:52:24.153352",
    "total_papers": 671
  }
}