May 2025 arXiv papers — page 20

Eric G. Stratman, Justin J. Boutilier, Laura A. Albert

Emergency Medical Services (EMS) in the United States and similar systems typically utilize a single treatment pathway, transporting all patients to emergency departments (EDs), regardless of their actual care needs or preferences. Recent policy reforms have sought to introduce alternative treatment pathways to divert lower acuity patients from the ED, but o

Kunlun Zhu, Jiaxun Zhang, Ziheng Qi, Nuoxing Shang

Recent advancements in large language model (LLM) agents have significantly accelerated scientific discovery automation, yet concurrently raised critical ethical and safety concerns. To systematically address these challenges, we introduce \textbf{SafeScientist}, an innovative AI scientist framework explicitly designed to enhance safety and ethical responsib

Xu Chu, Xinrong Chen, Guanyu Wang, Zhijie Tan

Inference time scaling drives extended reasoning to enhance the performance of Vision-Language Models (VLMs), thus forming powerful Vision-Language Reasoning Models (VLRMs). However, long reasoning dilutes visual tokens, causing visual information to receive less attention and may trigger hallucinations. Although introducing text-only reflection processes sh

Marc Jourdan, Gizem Yüce, Nicolas Flammarion

Recent advances in language modeling have underscored the role of preference feedback in enhancing model performance. This paper investigates the conditions under which preference feedback improves parameter estimation in classes of continuous parametric distributions. In our framework, the learner observes pairs of samples from an unknown distribution along

Wei Jie Yeo, Nirmalendu Prakash, Clement Neo, Roy Ka-Wei Lee

Refusal is a key safety behavior in aligned language models, yet the internal mechanisms driving refusals remain opaque. In this work, we conduct a mechanistic study of refusal in instruction-tuned LLMs using sparse autoencoders to identify latent features that causally mediate refusal behaviors. We apply our method to two open-source chat models and interve

Yanzhao Hou, Jiaxiang Geng, Boyu Li, Xiaofeng Tao

Federated LoRA has emerged as a promising technique for efficiently fine-tuning large language models (LLMs) on distributed devices by reducing the number of trainable parameters. However, existing approaches often inadequately overlook the theoretical and practical implications of system and data heterogeneity, thereby failing to optimize the overall traini

Hayden Moore, Sirui Qi, Ninad Hogade, Dejan Milojicic

In recent years, Large Language Models (LLM) such as ChatGPT, CoPilot, and Gemini have been widely adopted in different areas. As the use of LLMs continues to grow, many efforts have focused on reducing the massive training overheads of these models. But it is the environmental impact of handling user requests to LLMs that is increasingly becoming a concern.

Georgios Alexandris, Panagiotis Chaidos, Alexis Maras, Barry de Bruin

The ever-increasing complexity and operational diversity of modern Neural Networks (NNs) have caused the need for low-power and, at the same time, high-performance edge devices for AI applications. Coarse Grained Reconfigurable Architectures (CGRAs) form a promising design paradigm to address these challenges, delivering a close-to-ASIC performance while all

Alex Adams

This paper investigates the comparative performance of two fundamental approaches to solving linear regression problems: the closed-form Moore-Penrose pseudoinverse and the iterative gradient descent method. Linear regression is a cornerstone of predictive modeling, and the choice of solver can significantly impact efficiency and accuracy. I review and discu

Anthony Englert, Ian Dell'Antonio, Mireia Montes

Intracluster light, the diffuse glow of stars stripped from galaxies during a cluster's formation, is an established tracer of a cluster's dynamical history. The upcoming Vera C. Rubin Observatory's Legacy Survey of Space and Time (LSST) is set to revolutionize studies of intracluster light by imaging the entire southern sky down to a limiting surface bright

Patrick Achenbach, Daniel S. Carman, Ralf W. Gothe, Kyungseon Joo

Developing an understanding of phenomena driven by the emergence of hadron mass (EHM) is one of the most challenging problems in the Standard Model. This discussion focuses on the impact of results on nucleon resonance ($N^\ast$) electroexcitation amplitudes (or $\gamma_vpN^\ast$ electrocouplings) obtained from experiments during the 6-GeV era in Hall~B at J

Khashayar Etemadi, Marjan Sirjani, Mahshid Helali Moghadam, Per Strandberg

Cyber-physical systems (CPSs) are complex systems that integrate physical, computational, and communication subsystems. The heterogeneous nature of these systems makes their safety assurance challenging. In this paper, we propose a novel automated approach for guardrailing cyber-physical systems using property-based tests (PBTs) generated by Large Language M

Yuri Balashov

Large Language Models (LLMs) excel in translation among other things, demonstrating competitive performance for many language pairs in zero- and few-shot settings. But unlike dedicated neural machine translation models, LLMs are not trained on any translation-related objective. What explains their remarkable translation abilities? Are these abilities grounde

V. V. Samsonov, K. V. Nikolaev, B. I. Ostrovskii, S. N. Yakunin

We study X-ray diffraction in smectic liquid crystal multilayers. Such systems are fabricated as freely suspended films and have a unique layered structure. As such, they can be described as organic Bragg mirrors with sub-nanometer roughness. However, an interesting peculiarity arises in the diffraction on these structures: the characteristic shape of diffra

Yanqiu Ruan, Karthyek Murthy, Karthik Natarajan

We study decision-making problems where data comprises points from a collection of binary polytopes, capturing aggregate information stemming from various combinatorial selection environments. We propose a nonparametric approach for counterfactual inference in this setting based on a representative agent model, where the available data is viewed as arising f

Patrick Guidotti, Christoph Walker

In this paper a reduced one-dimensional moving boundary model is studied that describes the evolution of a biofilm driven by the presence of a reaction limiting substrate. Global well-posedness is established for the resulting parabolic free boundary value problem in strong form in Sobolev spaces and for a quasi-stationary approximation in spaces of classica

Elias Taira, Claire Kopenhafer, Brian W. O'Shea

In developing a deeper understanding of the Circumgalactic Medium, one feature that is poorly understood is the nature of the ultraviolet background (UVB) and its impact on observed column densities. A wide array of UVB models have been created over the years by many different authors, each based on the latest observational data available at the time. In add

Jan Ignatowicz, Krzysztof Kutt, Grzegorz J. Nalepa

The digitization of cultural heritage collections has opened new directions for research, yet the lack of enriched metadata poses a substantial challenge to accessibility, interoperability, and cross-institutional collaboration. In several past years neural networks models such as YOLOv11 and Detectron2 have revolutionized visual data analysis, but their app

Jonas E. Arias, Juan F. Rubio-Ramírez, Daniel Rudolf, Minchul Shin

We develop a new algorithm for inference in structural vector autoregressions (SVARs) identified with sign restrictions that can accommodate big data and modern identification schemes. The key innovation of our approach is to move beyond the traditional accept-reject framework commonly used in sign-identified SVARs. We show that an elliptical slice within Gi

Nada Cvetković, Han Cheng Lie

The work of Sprungk (Inverse Problems, 2020) established the local Lipschitz continuity of the misfit-to-posterior and prior-to-posterior maps with respect to the Kullback--Leibler divergence and the total variation, Hellinger, and 1-Wasserstein metrics, by proving certain upper bounds. The upper bounds were also used to show that if a posterior measure is m

Yunqiao Yang, Houxing Ren, Zimu Lu, Ke Wang

Recent advances in preference optimization have demonstrated significant potential for improving mathematical reasoning capabilities in large language models (LLMs). While current approaches leverage high-quality pairwise preference data through outcome-based criteria like answer correctness or consistency, they fundamentally neglect the internal logical coh

Shivani Chiranjeevi, Hossein Zaremehrjerdi, Zi K. Deng, Talukder Z. Jubery

The rapid global loss of biodiversity, particularly among insects, represents an urgent ecological crisis. Current methods for insect species discovery are manual, slow, and severely constrained by taxonomic expertise, hindering timely conservation actions. We introduce TerraIncognita, a dynamic benchmark designed to evaluate state-of-the-art multimodal mode

Polad Geidarov

Neural networks based on metric recognition methods have a strictly determined architecture. Number of neurons, connections, as well as weights and thresholds values are calculated analytically, based on the initial conditions of tasks: number of recognizable classes, number of samples, metric expressions used. This paper discusses the possibility of transfo

Kuntal Bhandari, Bingkang Huang, Šárka Nečasová

This paper is concerned with an interaction problem between a full compressible, electrically conducting fluid and a thermoelastic shell in a two-dimensional setting. The shell is modelled by linear thermoelasticity equations, and encompasses a time-dependent domain which is filled with a fluid described by full compressible (non-resistive) magnetohydrodynam

Nawar Turk, Eeham Khan, Leila Kosseim

This paper presents our approach to the SemEval-2025 Task~6 (PromiseEval), which focuses on verifying promises in corporate ESG (Environmental, Social, and Governance) reports. We explore three model architectures to address the four subtasks of promise identification, supporting evidence assessment, clarity evaluation, and verification timing. Our first mod

Giorgos Iacovides, Wuyang Zhou, Chao Li, Qibin Zhao

Tensor networks (TNs) provide efficient representations of high-dimensional data, yet identification of the optimal TN structures, the so called tensor network structure search (TN-SS) problem, remains a challenge. Current state-of-the-art (SOTA) algorithms solve TN-SS as a purely numerical optimization problem and require extensive function evaluations, whi

Wangting Zhou, Jiangshan He, Tong Cai, Lin Wang

Photoacoustic microscopy holds the potential to measure biomarkers' structural and functional status without labels, which significantly aids in comprehending pathophysiological conditions in biomedical research. However, conventional optical-resolution photoacoustic microscopy (OR-PAM) is hindered by a limited depth-of-field (DoF) due to the narrow depth ra

Afrozah Nadeem, Mark Dras, Usman Naseem

Large Language Models (LLMs) increasingly shape public discourse, yet most evaluations of political and economic bias have focused on high-resource, Western languages and contexts. This leaves critical blind spots in low-resource, multilingual regions such as Pakistan, where linguistic identity is closely tied to political, religious, and regional ideologies

Janik-Vasily Benzin, Gyunam Park, Stefanie Rinderle-Ma

Model abstraction (MA) and event abstraction (EA) are means to reduce complexity of (discovered) models and event data. Imagine a process intelligence project that aims to analyze a model discovered from event data which is further abstracted, possibly multiple times, to reach optimality goals, e.g., reducing model size. So far, after discovering the model,

Zhao Chen, Chen Shi, Christina Dan Wang

This paper investigates the estimation of the double autoregressive (DAR) model in the presence of skewed and heavy-tailed innovations. We propose a novel Normal Mixture Quasi-Maximum Likelihood Estimation (NM-QMLE) method to address the limitations of conventional quasi-maximum likelihood estimation (QMLE) under non-Gaussian conditions. By incorporating a n

Folco Giorgetti, Francesco Crocetti, Mario Luca Fravolini, Francesco Ferrante

In this paper, we address the problem of designing an aperiodic sampled-data controller stabilizing the zero-input equilibrium of an uncertain affine plant. The closed-loop system is modeled as a hybrid dynamical system incorporating a timer triggering the occurrence of the sampling events and two memory states storing the value of the controller state and c

Christoph Flathmann, Ulrich Ross, Jürgen Belz, Andreas Beyer

Momentum-resolved scanning transmission electron microscopy (MRSTEM) is a powerful phase-contrast technique that can map lateral magnetic and electric fields ranging from the micrometer to the subatomic scale. Resolving fields ranging from a few nanometers to a few hundred nanometers, as well as across material junctions, is particularly important since thes

Matthias Meister, Gabriel Müller, Patrick Boegel, Albert Roura

Atom interferometers deployed in space are excellent tools for high precision measurements, navigation, or Earth observation. In particular, differential interferometric setups feature common-mode noise suppression and enable reliable measurements in the presence of ambient platform noise. Here we report on orbital magnetometry campaigns performed with diffe

Ana Pavlaković

We study the stable pair theory on toric surfaces and determine the virtual tangent space over the fixed point loci. Further, we present a program to compute the virtual Euler characteristic, illustrated by the case of the projective plane. As an application, conjectures regarding rationality and symmetry are supported by verification of a special case.

LHCb collaboration, R. Aaij, A. S. W. Abdelmotteleb, C. Abellan Beteta

The substructure of jets in quantum chromodynamics (QCD) has garnered significant attention with the advent of infrared- and collinear-safe clustering algorithms and observables. A key question emerging from these studies is how in-jet emissions at soft and hard energy scales, across collinear and wide angles relative to the emitter, differ with the mass of

Shifeng Xie, Aref Einizade, Jhony H. Giraldo

Graph Representation Learning (GRL) is a fundamental task in machine learning, aiming to encode high-dimensional graph-structured data into low-dimensional vectors. Self-Supervised Learning (SSL) methods are widely used in GRL because they can avoid expensive human annotation. In this work, we propose a novel Subgraph Gaussian Embedding Contrast (SubGEC) met

Maria Eleftheria Vlontzou, Maria Athanasiou, Christos Davatzikos, Konstantina S. Nikita

The present study performs a comprehensive fairness analysis of machine learning (ML) models for the diagnosis of Mild Cognitive Impairment (MCI) and Alzheimer's disease (AD) from MRI-derived neuroimaging features. Biases associated with age, race, and gender in a multi-cohort dataset, as well as the influence of proxy features encoding these sensitive attri

Nathan Secrest, Sebastian von Hausegger, Mohamed Rameez, Roya Mohayaee

The Cosmological Principle, which states that the Universe is homogeneous and isotropic (when averaged on large scales), is the foundational assumption of Friedmann-Lemaitre-Robertson-Walker (FLRW) cosmologies such as the current standard Lambda-Cold-Dark-Matter ({\Lambda}CDM) model. This simplification yields an exact solution to the Einstein field equation

Jiahao Cui, Yan Chen, Mingwang Xu, Hanlin Shang

Generating highly dynamic and photorealistic portrait animations driven by audio and skeletal motion remains challenging due to the need for precise lip synchronization, natural facial expressions, and high-fidelity body motion dynamics. We propose a human-preference-aligned diffusion framework that addresses these challenges through two key innovations. Fir

Rui Xia, Dan Jiang, Quan Zhang, Ke Zhang

Temporal Action Localization (TAL) has garnered significant attention in information retrieval. Existing supervised or weakly supervised methods heavily rely on labeled temporal boundaries and action categories, which are labor-intensive and time-consuming. Consequently, unsupervised temporal action localization (UTAL) has gained popularity. However, current

Arjun Devraj, Eric Ding, Abhishek Vijaya Kumar, Robert Kleinberg

Distributed machine learning workloads use data and tensor parallelism for training and inference, both of which rely on the AllReduce collective to synchronize gradients or activations. However, AllReduce algorithms are delayed by the slowest GPU to reach the synchronization barrier before the collective (i.e., the straggler). To address this challenge, we

Fengxiang Wang, Mingshuo Chen, Xuming He, Yi-Fan Zhang

Existing benchmarks for multimodal learning in Earth science offer limited, siloed coverage of Earth's spheres and their cross-sphere interactions, typically restricting evaluation to the human-activity sphere of atmosphere and to at most 16 tasks. These limitations: narrow-source heterogeneity (single/few data sources), constrained scientific granularity, a

Mauro Pieroni

Primordial scalar curvature perturbations ($\zeta$), typically probed on large cosmological scales via CMB and LSS observations, can be significantly enhanced on smaller scales by various early Universe mechanisms, for instance, non-minimal inflationary models. While decoupled at linear order, scalar and tensor perturbations, i.e., Gravitational Waves (GWs),

Enrique Gaztanaga, K. Sravan Kumar, Swaraj Pradhan, Michael Gabler

We investigate the fully relativistic spherical collapse model of a uniform distribution of mass $M$ with initial comoving radius $\chi_*$ and spatial curvature $k \equiv 1/\chi_k^2 \le 1/\chi_*^2$ representing an over-density or bounded perturbation within a larger background. Our model incorporates a perfect fluid with an evolving equation of state, $P = P

Yu Zhang, Dong Guo, Fang Wu, Guoliang Zhu

Large Language Models (LLMs) with extended context lengths face significant computational challenges during the pre-filling phase, primarily due to the quadratic complexity of self-attention. Existing methods typically employ dynamic pattern matching and block-sparse low-level implementations. However, their reliance on local information for pattern identifi

Ruiqi He, Falk Lieder

People employ efficient planning strategies. But how are these strategies acquired? Previous research suggests that people can discover new planning strategies through learning from reinforcements, a process known as metacognitive reinforcement learning (MCRL). While prior work has shown that MCRL models can learn new planning strategies and explain more par

Hangoo Kang, Jehyeok Yeon, Gagandeep Singh

Autonomous agentic AI systems powered by vision-language models (VLMs) are rapidly advancing toward real-world deployment, yet their cross-modal reasoning capabilities introduce new attack surfaces for adversarial manipulation that exploit semantic reasoning across modalities. Existing adversarial attacks typically rely on visible pixel perturbations or requ

Simone Di Marino, Emanuele Naldi, Silvia Villa

This paper studies the convergence properties of the inexact Jordan-Kinderlehrer-Otto (JKO) scheme and proximal-gradient algorithm in the context of Wasserstein spaces. The JKO scheme, a widely-used method for approximating solutions to gradient flows in Wasserstein spaces, typically assumes exact solutions to iterative minimization problems. However, practi

Marco Hirsch, Peter Hevesi, Paul Lukowicz

We present CASE, an open-source framework for adaptive participatory research and disease surveillance. Unlike traditional survey platforms with static branching logic, CASE uses an event-driven architecture that adjusts survey workflows in real time based on participant responses, external data, temporal conditions, and evolving participant state. This desi

Sanberk Serbest, Tijana Stojkovic, Milos Cernak, Andrew Harper

In this work, we propose a full-band real-time speech enhancement system with GAN-based stochastic regeneration. Predictive models focus on estimating the mean of the target distribution, whereas generative models aim to learn the full distribution. This behavior of predictive models may lead to over-suppression, i.e. the removal of speech content. In the li

Ellyn K. Baines, James H. Clark, Henrique R. Schmitt, Jordan M. Stone

We present new angular diameter measurements for 33 stars from the Navy Precision Optical Interferometer, reaching uncertainties on the limb-darkened diameter of 2% or less for 21 targets. We also determined the physical radius, bolometric flux, luminosity, and effective temperature for each star. Our sample is a mix of giant, subgiant, and dwarf stars, and

Johannes Buchner

The goal of these notes is to make the concept of "pseudo goodwin cycles" mathematically more precise. At first the title seems like a contradiction to have a wage-led model and still find goodwin cycles in it, but the point we try to make in the paper is that those are only `pseudo-goodwin' cycles, and not real goodwin cycles.

Jingfu Zhang, Swathi S. Hegde, Fedor Jelezko, Dieter Suter

Long coherence times rank among the most important performance measures for many different types of quantum technology. In NV centers of diamond, the nuclear spins provide particularly long dephasing times. However, since initialization and readout require assistance from the electron spin, the apparent dephasing times can be reduced by the electron spin lif

Sang Hu, Zihan Zhou

This paper studies the dividend and capital injection problem under a diffusion risk model with general discount functions. A proportional cost is imposed when injecting capitals. For exponential discounting as time-consistent benchmark, we obtain the closed-form solutions and show that the optimal strategies are of threshold type. Under general discount fun

Stepan Trifonov, Leonid Levin, Savelii Chezhegov, Aleksandr Beznosikov

Machine learning and deep learning are widely researched fields that provide solutions to many modern problems. Due to the complexity of new problems related to the size of datasets, efficient approaches are obligatory. In optimization theory, the Heavy Ball and Nesterov methods use \textit{momentum} in their updates of model weights. On the other hand, the

Andrew Chang, Yike Li, Iran R. Roman, David Poeppel

Audio DNNs have demonstrated impressive performance on various machine listening tasks; however, most of their representations are computationally costly and uninterpretable, leaving room for optimization. Here, we propose a novel approach centered on spectrotemporal modulation (STM) features, a signal processing method that mimics the neurophysiological rep

Rebecca Ramnauth, Dražen Brščić, Brian Scassellati

From dating to job interviews, making new friends or simply chatting with the cashier at checkout, engaging in small talk is a vital, everyday social skill. For adults with Autism Spectrum Disorder (ASD), small talk can be particularly challenging, yet it is essential for social integration, building relationships, and accessing professional opportunities. I

Toshiyuki Akita, Kakeru Shikata

In this paper, we investigate the structure of associated groups of symmetric quandles. Among other results, we explore the relationship between the associated group of a symmetric quandle and that of its underlying quandle. We provide a group-theoretic characterization of associated groups of symmetric quandles. Furthermore, we show that a symmetric quandle

Sebastián Jiménez, Mira Jürgens, Willem Waegeman

Identifying and disentangling sources of predictive uncertainty is essential for trustworthy supervised learning. We argue that widely used second-order methods that disentangle aleatoric and epistemic uncertainty are fundamentally incomplete. First, we show that unaccounted bias contaminates uncertainty estimates by overestimating aleatoric (data-related) u

Jan Ignatowicz, Krzysztof Kutt, Grzegorz J. Nalepa

Digitizing cultural heritage collections has become crucial for preservation of historical artifacts and enhancing their availability to the wider public. Galleries, libraries, archives and museums (GLAM institutions) are actively digitizing their holdings and creates extensive digital collections. Those collections are often enriched with metadata describin

Masaki Murooka, Iori Kumagai, Mitsuharu Morisawa, Fumio Kanehiro

In this letter, we propose an efficient and highly versatile loco-manipulation planning for humanoid robots. Loco-manipulation planning is a key technological brick enabling humanoid robots to autonomously perform object transportation by manipulating them. We formulate planning of the alternation and sequencing of footsteps and grasps as a graph search prob

Liyun Zhu, Qixiang Chen, Xi Shen, Xiaodong Cun

Video Anomaly Understanding (VAU) is essential for applications such as smart cities, security surveillance, and disaster alert systems, yet remains challenging due to its demand for fine-grained spatio-temporal perception and robust reasoning under ambiguity. Despite advances in anomaly detection, existing methods often lack interpretability and struggle to

Shibbir Ahmed, Shahnewaz Karim Sakib, Anindya Bijoy Das

This study presents a multimodal AI framework designed for precisely classifying medical diagnostic images. Utilizing publicly available datasets, the proposed system compares the strengths of convolutional neural networks (CNNs) and different large language models (LLMs). This in-depth comparative analysis highlights key differences in diagnostic performanc

Mingtai Xie, Zheng Zhang, Weizhen Zhuo, Wei Xu

Realizing Kitaev interactions on triangular lattices offers a compelling platform for exploring quantum-spin-liquid physics beyond the conventional honeycomb lattice framework. Here, we investigate the triangular-lattice antiferromagnet KCeSe2, where multiple probes reveal strong magnetic anisotropy suggesting significant Kitaev physics. Through detailed and

Masaki Murooka, Kei Okada, Masayuki Inaba

Whole-body contact is an effective strategy for improving the stability and efficiency of the motion of robots. For robots to automatically perform such motions, we propose a posture generation method that employs all available surfaces of the robot links. By representing the contact point on the body surface by two-dimensional configuration variables, the j

Eva Martín del Pico, Josep Lluís Gelpí, Salvador Capella-Gutiérrez

Software is an essential component of research. However, little attention has been paid to it compared with that paid to research data. Recently, there has been an increase in efforts to acknowledge and highlight the importance of software in research activities. Structured metadata from platforms like bio.tools, Bioconductor, and Galaxy ToolShed offers valu

Masaki Murooka, Mitsuharu Morisawa, Fumio Kanehiro

Multi-contact motion is important for humanoid robots to work in various environments. We propose a centroidal online trajectory generation and stabilization control for humanoid dynamic multi-contact motion. The proposed method features the drastic reduction of the computational cost by using preview control instead of the conventional model predictive cont

Amanda Quirk, Tom Rice

Universal Design (UD), an approach to accessibility that was first conceptualized in architecture to make buildings physically accessible, has since been applied to curriculum design to make classrooms accessible for a larger range of learning needs. In this paper, we illustrate how the concepts of UD are relevant outside of architecture and the creation of

Tomoyoshi Ibukiyama, Brandon Williams

We state conjectures that relate Hermitian modular forms of degree two and algebraic modular forms for the compact group $SO(6)$. We provide evidence for these conjectures in the form of dimension formulas and explicit computations of eigenforms.

Sabina J. Sloman, Michele Caprio, Samuel Kaski

Uncertainty-aware machine learners, such as Bayesian neural networks, output a quantification of uncertainty instead of a point prediction. We provide uncertainty-aware learners with a principled framework to characterize, and identify ways to eliminate, errors that arise from reducible (epistemic) uncertainty. We introduce a principled definition of epistem

Liangliang Zhang, Zhuorui Jiang, Hongliang Chi, Haoyang Chen

Knowledge Graph Question Answering (KGQA) systems rely on high-quality benchmarks to evaluate complex multi-hop reasoning. However, despite their widespread use, popular datasets such as WebQSP and CWQ suffer from critical quality issues, including inaccurate or incomplete ground-truth annotations, poorly constructed questions that are ambiguous, trivial, or

Nicol Visser, Herman Kamper

Spoken language models (SLMs) operate on acoustic units obtained by discretizing self-supervised speech representations. Although the characteristics of these units directly affect performance, the interaction between codebook size and unit coarseness (i.e., duration) remains unexplored. We investigate SLM performance as we vary codebook size and unit coarse

Kaijie Chen, Zihao Lin, Zhiyang Xu, Ying Shen

Reasoning is a fundamental capability often required in real-world text-to-image (T2I) generation, e.g., generating ``a bitten apple that has been left in the air for more than a week`` necessitates understanding temporal decay and commonsense concepts. While recent T2I models have made impressive progress in producing photorealistic images, their reasoning

Maxime Gonthier, Dante D. Sanchez-Gallegos, Haochen Pan, Bogdan Nicolae

The exponential growth of data necessitates distributed storage models, such as peer-to-peer systems and data federations. While distributed storage can reduce costs and increase reliability, the heterogeneity in storage capacity, I/O performance, and failure rates of storage resources makes their efficient use a challenge. Further, node failures are common

Shubham Jain, Surjit Kumar, Milan Kumar Mal, Paramita Pramanick

We introduce a Schur-Agler type class associated with the tetrablock and establish a realization theorem for this class. Furthermore, we provide a tetrablock analog of the interpolation theorem, extension theorem, and the Toeplitz corona theorem.

Sándor Lökös

Recent theoretical results renewed the interest in charged particle multiplicity distributions. The Shannon entropy of such distributions is conjectured to be related to the entanglement or von Neumann entropy of partonic quantum system. In this paper, we show that the measured charged particle multiplicities can be derived from the principle of maximum entr

Ildus Sadrtdinov, Ivan Klimov, Ekaterina Lobacheva, Dmitry Vetrov

We present a thermodynamic interpretation of the stationary behavior of stochastic gradient descent (SGD) under fixed learning rates (LRs) in neural network training. We show that SGD implicitly minimizes a free energy function $F=U-TS$, balancing training loss $U$ and the entropy of the weights distribution $S$, with temperature $T$ determined by the LR. Th

Guillermo Barajas

Let $X$ be a compact Riemann surface, $\Gamma$ a finite group of automorphisms of $X$ and $G$ a connected reductive complex Lie group with center $Z$. If we equip this data with a homomorphism $\theta:\Gamma\to\text{Aut}(G)$ and a 2-cocycle $c:\Gamma\times\Gamma\to Z$, there is a notion of $(\theta,c)$-twisted $\Gamma$-equivariant $G$-bundle over $X$. The ai

Justin Cammarota, Jian-Wei Qiu, Kazuhiro Watanabe, Jia-Yue Zhang

We present the first calculation of next-to-leading order (NLO) factorized QED and QCD contributions to the short-distance hard coefficients of inclusive lepton-hadron deep inelastic scattering (DIS) in a joint QED and QCD factorization approach. Unlike the traditional radiative correction approach to handle the collision-induced QED contributions to DIS, QE

Ke Weng, Lun Du, Sirui Li, Wangyue Lu

Autoformalization, the process of transforming informal mathematical propositions into verifiable formal representations, is a foundational task in automated theorem proving, offering a new perspective on the use of mathematics in both theoretical and applied domains. Driven by the rapid progress in artificial intelligence, particularly large language models

Shi-Xue Zhang, Hongfa Wang, Duojun Huang, Xin Li

Video captions play a crucial role in text-to-video generation tasks, as their quality directly influences the semantic coherence and visual fidelity of the generated videos. Although large vision-language models (VLMs) have demonstrated significant potential in caption generation, existing benchmarks inadequately address fine-grained evaluation, particularl

M. Gorgone, C. F. Munafo', A. Palumbo, P. Rogolino

A complete thermodynamical analysis for a blood model, based on mixture theory, is performed. The model is developed considering the blood as a suspension of red blood cells (solid component) in the plasma (fluid component), and taking into account the temperature effects. Furthermore, two independent scalar internal variables are introduced accounting for a

Peisen Yuan, Beatriz Martín-García, Evgeny Modin, M. Xochitl Aguilar-Pujol

Nonlocal magnon transport can provide valuable insight into the magnetic properties of magnetic insulators (MIs). A spin-flop transition, a typical magnetic reorientation in antiferromagnets, is expected to affect mag non transport, but studies on this topic are still rare and remain challenging, especially for van der Waals materials. Here we demonstrate th

Polad Geidarov

The paper discusses the capabilities of multilayer perceptron neural networks implementing metric recognition methods, for which the values of the weights are calculated analytically by formulas. Comparative experiments in training a neural network with pre-calculated weights and with random initialization of weights on different sizes of the MNIST training

Mohamed Rayan Barhdadi, Hasan Kurban, Hussein Alnuweiri

PhysicsNeRF is a physically grounded framework for 3D reconstruction from sparse views, extending Neural Radiance Fields with four complementary constraints: depth ranking, RegNeRF-style consistency, sparsity priors, and cross-view alignment. While standard NeRFs fail under sparse supervision, PhysicsNeRF employs a compact 0.67M-parameter architecture and ac

Keqin Peng, Liang Ding, Yuanxin Ouyang, Meng Fang

Reasoning Large Language Models (RLLMs) have demonstrated impressive performance on complex tasks, largely due to the adoption of Long Chain-of-Thought (Long CoT) reasoning. However, they often exhibit overthinking -- performing unnecessary reasoning steps even after arriving at the correct answer. Prior work has largely focused on qualitative analyses of ov

Haiqi Gao, Yu Shao, Jiaming Liang, Xuehui Wang

Optical encryption inherently provides strong security advantages, with hybrid optoelectronic systems offering additional degrees of freedom by integrating optical and algorithmic domains. However, existing optical encryption schemes heavily rely on electronic computation, limiting overall efficiency, while the physical keys are susceptible to damage, compro

Andrea Merlo, Mihalis Mourgoglou, Carmelo Puliatti

For $n \geq 2$, we consider the operator $L_A = -\mathrm{div }(A(\cdot)\nabla)$, where $A$ is a uniformly elliptic $(n+1)\times(n+1)$ matrix with variable coefficients, a Radon measure $\mu$ on $\mathbb{R}^{n+1}$, and the associated gradient of the single layer potential operator $T_\mu$. Under a Dini-type assumption on the mean oscillation of the matrix $A$

Krithik Vishwanath, Anton Alyakin, Mrigayu Ghosh, Jin Vivian Lee

The Congress of Neurological Surgeons Self-Assessment for Neurological Surgeons (CNS-SANS) questions are widely used by neurosurgical residents to prepare for written board examinations. Recently, these questions have also served as benchmarks for evaluating large language models' (LLMs) neurosurgical knowledge. This study aims to assess the performance of s

Nuno Brito, Manuel Colaço, Orlando Oliveira, Paulo J. Silva

The computation of the four-gluon and ghost-gluon vertices in the Landau gauge using high statistical lattice ensembles for $32^4$ and $48^4$ volumes is addressed. For the four-gluon vertex, our previous results for the collinear kinematics are updated allowing to get a better coverage of the IR region. Furthermore, the one-particle irreducible ghost-gluon G

Ron Shapira Weber, Shahar Ben Ishay, Andrey Lavrinenko, Shahaf E. Finder

Fast and scalable alignment of time series is a fundamental challenge in many domains. The standard solution, Dynamic Time Warping (DTW), struggles with poor scalability and sensitivity to noise. We introduce TimePoint, a self-supervised method that dramatically accelerates DTW-based alignment while typically improving alignment accuracy by learning keypoint

Xiang Li, Haiyang Yu, Xinghua Zhang, Ziyang Huang

Process Reward Models (PRMs) are crucial in complex reasoning and problem-solving tasks (e.g., LLM agents with long-horizon decision-making) by verifying the correctness of each intermediate reasoning step. In real-world scenarios, LLMs may apply various reasoning patterns (e.g., decomposition) to solve a problem, potentially suffering from errors under vari

Xiaorui Wu, Fei Li, Xiaofeng Mao, Xin Zhang

Large language models (LLMs) frequently refuse to respond to pseudo-malicious instructions: semantically harmless input queries triggering unnecessary LLM refusals due to conservative safety alignment, significantly impairing user experience. Collecting such instructions is crucial for evaluating and mitigating over-refusals, but existing instruction curatio

Jonathan Smith, Siddartha Khastgir

If public trust is lost in a new technology early in its life cycle it can take much more time for the benefits of that technology to be realised. Eventually tens-of-millions of people will collectively have the power to determine self-driving technology success of failure driven by their perception of risk, data handling, safety, governance, accountability,

Jun Yang, Cheng-Chi Wang, Bogdan Alexandru Stoica, Kexin Pei

Large Language Models (LLMs) have been increasingly used to optimize code efficiency. Evaluating their effectiveness and further suggesting optimization opportunities often rely on high-quality tests to demonstrate the performance bottlenecks presented in the program. However, existing approaches rely on a limited set of hand-curated inputs or LLM-generated

Chenjie Li, Amir Gilad, Boris Glavic, Zhengjie Miao

Programmatic weak supervision (PWS) significantly reduces human effort for labeling data by combining the outputs of user-provided labeling functions (LFs) on unlabeled datapoints. However, the quality of the generated labels depends directly on the accuracy of the LFs. In this work, we study the problem of fixing LFs based on a small set of labeled examples

Prachi Jadhav, Hongwei Jin, Ewa Deelman, Prasanna Balaprakash

High-Performance Computing (HPC) job scheduling involves balancing conflicting objectives such as minimizing makespan, reducing wait times, optimizing resource use, and ensuring fairness. Traditional methods, including heuristic-based, e.g., First-Come-First-Served (FJFS) and Shortest Job First (SJF), or intensive optimization techniques, often lack adaptabi

Zhuodong Li, Fei Hou, Wencheng Wang, Xuequan Lu

Orienting point clouds is a fundamental problem in computer graphics and 3D vision, with applications in reconstruction, segmentation, and analysis. While significant progress has been made, existing approaches mainly focus on watertight, object-level 3D models. The orientation of large-scale, non-watertight 3D scenes remains an underexplored challenge. To a

Leonard Edens, Francisco Romero Lara, Trisha Sai, Kalyan Biswas

Tailor-made graphene nanostructures can exhibit symmetry-protected topological boundary states that host localized spin-$1/2$ moments. However, one frequently observes charge transfer on coinage metal substrates, which results in spinless closed-shell configurations. Using low temperature scanning tunneling spectroscopy, we demonstrate here that pristine top

Adolfo Cisterna, Mokhtar Hassaine, Ulises Hernandez-Vera

We study the thermodynamics of a class of four-dimensional black hole solutions arising from the compactification of a higher-curvature gravity theory featuring an infinite tower of Lovelock-type invariants. For planar horizons, we identify two distinct branches: a regular black hole supported by a nontrivial scalar field and a non-regular general relativity