May 2025 arXiv papers — page 108

Ben Anson, Xi Wang, Laurence Aitchison

One persistent challenge in LLM research is the development of attention mechanisms that are able to generalise from training on shorter contexts to inference on longer contexts. We propose two conditions that we expect all effective long context attention mechanisms to have: scale-invariant total attention, and scale-invariant attention sparsity. Under a Ga

Dongkeun Yoon, Seungone Kim, Sohee Yang, Sunkyoung Kim

Despite their strengths, large language models (LLMs) often fail to communicate their confidence accurately, making it difficult to assess when they might be wrong and limiting their reliability. In this work, we demonstrate that reasoning models that engage in extended chain-of-thought (CoT) reasoning exhibit superior performance not only in problem-solving

Gyu Rac Lee, Thomas Defferriere, Jinwook Kim, Han Gil Seo

Although strong modulation of interfacial electron concentrations by the relative acidity of surface additives has been suggested, direct observation of corresponding changes in surface conductivity, crucial for understanding the role of local space charge, has been lacking. Here, we introduce a model platform comprising well-aligned mixed ionic-electronic c

Yuya Ominato, Masaki Yama, Ai Yamakage, Mamoru Matsuo

In this review, we present recent theoretical developments on spin transport phenomena probed by ferromagnetic resonance (FMR) modulation in two-dimensional systems coupled to magnetic materials. We first address FMR linewidth enhancements induced by spin pumping at interfaces, emphasizing their potential as sensitive probes of superconducting pairing symmet

Mahdi Hejrati, Pauli Mustalahti, Jouni Mattila

This paper presents an immersive bilateral teleoperation framework for beyond-human-scale manipulators, combining motion/force transparency with enhanced operator embodiment through virtual reality (VR) and distributed haptic feedback. The platform integrates a full-scale industrial hydraulic manipulator, a 7-DoF haptic exoskeleton, and head-tracked visual f

Rupam Samanta, Wojciech Broniowski

We study magnetic properties of the Hadron Resonance Gas in the presence of a strong ($0 \le B \le 0.15~{\rm GeV}^2$) uniform magnetic field, using physical values of the magnetic moments of hadrons, i.e., including their anomalous parts. The values of these moments are taken from experiment, or when unavailable, from theoretical estimates. We evaluate the c

Agam Goyal, Xianyang Zhan, Yilun Chen, Koustuv Saha

Large language models (LLMs) have shown great potential in flagging harmful content in online communities. Yet, existing approaches for moderation require a separate model for every community and are opaque in their decision-making, limiting real-world adoption. We introduce Mixture of Moderation Experts (MoMoE), a modular, cross-community framework that add

Pedro H. Azevedo de Amorim, Satoshi Kura, Philip Saville

We give a denotational account of logical relations for call-by-push-value (CBPV) in the fibrational style of Hermida, Jacobs, Katsumata and others. Fibrations -- which axiomatise the usual notion of sets-with-relations -- provide a clean framework for constructing new, logical relations-style, models. Such models can then be used to study properties such as

He Zhu, Junyou Su, Minxin Chen, Wen Wang

In the field of urban planning, existing Vision-Language Models (VLMs) frequently fail to effectively analyze and evaluate planning maps, despite the critical importance of these visual elements for urban planners and related educational contexts. Planning maps, which visualize land use, infrastructure layouts, and functional zoning, require specialized unde

Samrat Roy, Marina Bogomolov, Ruth Heller, Amy M. Claridge

The long term consequences of unwanted pregnancies carried to term on mothers have not been much explored. We use data from the Wisconsin Longitudinal Study (WLS) and propose a novel approach, namely two team cross-screening, to study the possible effects of unwanted pregnancies carried to term on various aspects of mothers' later-life mental health, physica

Simon Geirnaert, Simon L. Kappel, Preben Kidmose

Current assistive hearing devices, such as hearing aids and cochlear implants, lack the ability to adapt to the listener's focus of auditory attention, limiting their effectiveness in complex acoustic environments like cocktail party scenarios where multiple conversations occur simultaneously. Neuro-steered hearing devices aim to overcome this limitation by

Maria Panagiotou, Lorenzo Brigato, Vivien Streit, Amanda Hayoz

Despite recent advances in insulin preparations and technology, adjusting insulin remains an ongoing challenge for the majority of people with type 1 diabetes (T1D) and longstanding type 2 diabetes (T2D). In this study, we propose the Adaptive Basal-Bolus Advisor (ABBA), a personalised insulin treatment recommendation approach based on reinforcement learning

Farshad Sangari Abiz, Reshad Hosseini, Babak N. Araabi

Variational Autoencoders (VAEs) are powerful generative models for learning latent representations. Standard VAEs generate dispersed and unstructured latent spaces by utilizing all dimensions, which limits their interpretability, especially in high-dimensional spaces. To address this challenge, Variational Sparse Coding (VSC) introduces a spike-and-slab prio

David Damanik, Jake Fillman, Giorgio Young

We prove a dispersive estimate for periodic discrete Schr\"odinger operators on the line with optimal rate of decay. Additionally, by standard methods, we deduce dispersive estimates for the discrete nonlinear Schr\"odinger equation with small initial data and suitable nonlinearity when the underlying Hamiltonian is periodic.

Jiaxi Zha

In this work, we first study the cotensor product of comodules in the $\infty$-category $\mathrm{Mod}_R$ for a connected $\mathbb{E}_{\infty}$-ring spectrum $R$. We then apply these results to analyze higher coalgebra structures of topological coHochschild homology (coTHH) and establish its Morita-Takeuchi invariance, which are precisely dual to the correspo

Sribalaji C. Anand, Alexander J Gallo, Nicola Bastianello

Consensus algorithms are fundamental to multi-agent distributed optimization, and their security under adversarial conditions is an active area of research. While prior works primarily establish conditions for successful global consensus under attack, little is known about system behavior when these conditions are violated. This paper addresses this gap by i

Eduardo Camps Moreno, Adrián Fidalgo-Díaz, Hiram H. López, Umberto Martínez-Peñas

Multivariate multiplicity codes have been recently explored because of their importance for list decoding and local decoding. Given a multivariate multiplicity code, in this paper, we compute its dimension using Gr\"obner basis tools, its dual in terms of indicator functions, and explicitly describe a parity-check matrix. In contrast with Reed--Muller, Reed-

Tong Li, Jiachuan Wang, Yongqi Zhang, Shuangyin Li

Citation classification, which identifies the intention behind academic citations, is pivotal for scholarly analysis. Previous works suggest fine-tuning pretrained language models (PLMs) on citation classification datasets, reaping the reward of the linguistic knowledge they gained during pretraining. However, directly fine-tuning for citation classification

Nadav Har-Tuv, Or Tal, Yossi Adi

We present PAST, a novel end-to-end framework that jointly models phonetic information alongside signal reconstruction, eliminating the need for external pretrained models. Unlike previous approaches that rely on pretrained self-supervised models, PAST employs supervised phonetic data, directly integrating domain knowledge into the tokenization process via a

Somnath Banerjee, Pratyush Chatterjee, Shanu Kumar, Sayan Layek

While LLMs appear robustly safety-aligned in English, we uncover a catastrophic, overlooked weakness: attributional collapse under code-mixed perturbations. Our systematic evaluation of open models shows that the linguistic camouflage of code-mixing -- ``blending languages within a single conversation'' -- can cause safety guardrails to fail dramatically. At

Yifan Sui, Hao Wang, Hanfei Yu, Yitao Hu

Serverless computing has grown rapidly for serving Large Language Model (LLM) inference due to its pay-as-you-go pricing, fine-grained GPU usage, and rapid scaling. However, our analysis reveals that current serverless can effectively serve general LLM but fail with Low-Rank Adaptation (LoRA) inference due to three key limitations: 1) massive parameter redun

Mani Shemiranifar

Despite advances in transformer-based language models (LMs), a fundamental question remains largely unanswered: Are all layers activated during inference? We investigate this question by detecting unactivated layers (which we refer to as Voids) using a non-trainable and parameter-free adaptive computation method called L2 Adaptive Computation (LAC). We adapt

Tim C. Rese, Alexandra Kapp, David Bermbach

The growing number of moving Internet-of-Things (IoT) devices has led to a surge in moving object data, powering applications such as traffic routing, hotspot detection, or weather forecasting. When managing such data, spatial database systems offer various index options and data formats, e.g., point-based or trajectory-based. Likewise, dataset characteristi

Aviv Navon, Aviv Shamsian, Yael Segal-Feldman, Neta Glazer

Target speaker extraction (TSE) aims to isolate a specific speaker's speech from a mixture using speaker enrollment as a reference. While most existing approaches are discriminative, recent generative methods for TSE achieve strong results. However, generative methods for TSE remain underexplored, with most existing approaches relying on complex pipelines an

Xiaoyu Tian, Yunjie Ji, Haotian Wang, Shuaiting Chen

Distillation has emerged as a practical and effective approach to enhance the reasoning capabilities of open-source language models. In this work, we conduct a large-scale empirical study on reasoning data distillation by collecting verified outputs from three state-of-the-art teacher models-AM-Thinking-v1, Qwen3-235B-A22B, and DeepSeek-R1-on a shared corpus

Xinxin Fan, Wenxiong Chen, Mengfan Li, Wenqi Wei

Adversarial attacks to graph analytics are gaining increased attention. To date, two lines of countermeasures have been proposed to resist various graph adversarial attacks from the perspectives of either graph per se or graph neural networks. Nevertheless, a fundamental question lies in whether there exists an intrinsic adversarial resilience state within a

Jiaang Li, Yifei Yuan, Wenyan Li, Mohammad Aliannejadi

As vision-language models (VLMs) become increasingly integrated into daily life, the need for accurate visual culture understanding is becoming critical. Yet, these models frequently fall short in interpreting cultural nuances effectively. Prior work has demonstrated the effectiveness of retrieval-augmented generation (RAG) in enhancing cultural understandin

Mohammed Barhoush, Ryo Nishimaki, Takashi Yamakawa

We investigate two natural relaxations of quantum cryptographic primitives. The first involves quantum input sampling, where inputs are generated by a quantum algorithm rather than sampled uniformly at random. Applying this to pseudorandom generators ($\textsf{PRG}$s) and pseudorandom states ($\textsf{PRS}$s), leads to the notions denoted as $\textsf{PRG}^{q

Matjaž Omladič, Martin Vuk, Aljaž Zalar

A recent survey, nicknamed "Hitchhiker's Guide", J.J. Arias-Garc{\i}a, R. Mesiar, and B. De Baets, A hitchhiker's guide to quasi-copulas, Fuzzy Sets and Systems 393 (2020) 1-28, has raised the rating of quasi-copula problems in the dependence modeling community in spite of the lack of statistical interpretation of quasi-copulas. In our previous work (Fuzzy S

Tianhe Wu, Jian Zou, Jie Liang, Lei Zhang

DeepSeek-R1 has demonstrated remarkable effectiveness in incentivizing reasoning and generalization capabilities of large language models (LLMs) through reinforcement learning. Nevertheless, the potential of reasoning-induced computation has not been thoroughly explored in the context of image quality assessment (IQA), a task depending critically on visual r

Kamal Singh, Sami Marouani, Ahmad Al Sheikh, Pham Tran Anh Quang

Reinforcement learning (RL) has been increasingly applied to network control problems, such as load balancing. However, existing RL approaches often suffer from lack of interpretability and difficulty in extracting controller equations. In this paper, we propose the use of Kolmogorov-Arnold Networks (KAN) for interpretable RL in network control. We employ a

Imon Banerjee, Vinayak Rao, Harsha Honnappa

Estimating the transition dynamics of controlled Markov chains is crucial in fields such as time series analysis, reinforcement learning, and system exploration. Traditional non-parametric density estimation methods often assume independent samples and require oracle knowledge of smoothness parameters like the H\"older continuity coefficient. These assumptio

Andrey Alexandrov, Giovanni Acampora, Giovanni De Lellis, Antonia Di Crescenzo

Accurately tracking particles and determining their coordinate along the optical axis is a major challenge in optical microscopy, especially when extremely high precision is needed. In this study, we introduce a deep learning approach using convolutional neural networks (CNNs) that can determine axial coordinates from dual-focal-plane images without relying

Huayuan Huang, M. Kanat Camlibel, Raffaella Carloni, Henk J. van Waarde

In this study, we propose new global stabilization approaches for a class of polynomial systems in both model-based and data-driven settings. The existing model-based approach guarantees global asymptotic stability of the closed-loop system only when the Lyapunov function is radially unbounded, which limits its applicability. To overcome this limitation, we

O. A. Dobush, M. P. Kozlovskii, I. V. Pylyuk, Yu. O. Plevachuk

We investigate how varying two microscopic parameters - cell volume and the ratio between repulsion and attraction intensities - affect the phase behavior of a cell model with a Curie-Weiss-type interaction. The analysis is based on an exact solution previously derived for this model in the grand canonical ensemble. At sufficiently low temperatures, the cell

Chihan Huang, Hao Tang

Although autoregressive models have dominated language modeling in recent years, there has been a growing interest in exploring alternative paradigms to the conventional next-token prediction framework. Diffusion-based language models have emerged as a compelling alternative due to their powerful parallel generation capabilities and inherent editability. How

Xuyang Liu, Yiyu Wang, Junpeng Ma, Linfeng Zhang

Video large language models (VideoLLM) excel at video understanding, but face efficiency challenges due to the quadratic complexity of abundant visual tokens. Our systematic analysis of token compression methods for VideoLLMs reveals two critical issues: (i) overlooking distinctive visual signals across frames, leading to information loss; (ii) suffering fro

Xianghua Zeng, Hao Peng, Angsheng Li

Although Graph Neural Networks (GNNs) have shown promising potential in fake news detection, they remain highly vulnerable to adversarial manipulations within social networks. Existing methods primarily establish connections between malicious accounts and individual target news to investigate the vulnerability of graph-based detectors, while they neglect the

Lance T. Wilhelm, Xiaohan Ding, Kirk McInnis Knutsen, Buse Carik

Effective workplace communication is essential for managerial success, yet many managers lack access to tailored and sustained training. Although AI-assisted communication systems may offer scalable training solutions, little is known about how managers envision the role of AI in helping them improve their communication skills. To investigate this, we design

Md Atik Ahamed, Qiang Ye, Qiang Cheng

Missing values in high-dimensional, mixed-type datasets pose significant challenges for data imputation, particularly under Missing Not At Random (MNAR) mechanisms. Existing methods struggle to integrate local and global data characteristics, limiting performance in MNAR and high-dimensional settings. We propose an innovative framework, RefiDiff, combining l

Daiki Sasaki, Ryosuke Koga, Taihei Kuroiwa, Yuya Ito

We propose a Hamiltonian-level framework for non-Markovian quantum reservoir computing directly tailored for analog hardware implementations. By dividing the reservoir into a system block and an environment block and evolving their joint state under a unified Hamiltonian, our architecture naturally embeds memory backflow by harnessing entanglement-induced in

Yi-Cheng Lin, Huang-Cheng Chou, Hung-yi Lee

While subgroup disparities and performance bias are increasingly studied in computational research, fairness in categorical Speech Emotion Recognition (SER) remains underexplored. Existing methods often rely on explicit demographic labels, which are difficult to obtain due to privacy concerns. To address this limitation, we introduce an Implicit Demography I

Igor Lugo, Martha G. Alatriste-Contreras, Rafael Sánchez-Guevara

Audio signals in a set of musical pieces are modeled as a complex network for studying the relationship between the complexity of frequency fluctuations and the interpretive style of the bass viola da gamba. Based on interdisciplinary scientific and music approaches, we compute the spectral decomposition and translated its frequency components to a network o

The LHAASO Collaboration, Zhen Cao, F. Aharonian, Y. X. Bai

We report the high-purity identification of cosmic-ray (CR) protons and a precise measurement of their energy spectrum from 0.15 to 12 PeV using the Large High Altitude Air Shower Observatory (LHAASO). Abundant event statistics, combined with the simultaneous detection of electrons/photons, muons, and Cherenkov light in air showers, enable spectroscopic meas

Mikko Partanen, Jukka Tulkki

Light does not travel in a perfectly straight line when it passes near massive objects. In this work, we calculate the gravitational deflection of light using the gauge theory of unified gravity [Rep. Prog. Phys. 88, 057802 (2025)], formulated as an extension of the Standard Model. The nonlinear graviton-graviton interaction is accounted for in the lowest or

Aaron Bertram, Brooke Ullery

The projective space of symmetric tensors of degree d can be reinterpreted as a projective space of finite, graded Gorenstein rings with socle in degree d. Via a pair of explicit stability conditions (one for even values of d and one for odd values), the space of symmetric tensors is partitioned by Harder-Narasimhan filtration type. This is worked out explic

Samuel Boissiere, Marc Nieper-Wisskirchen, Gregory Sankaran

We construct a birational model of the generalised Kummer fourfold of the Jacobian of a genus two curve, based on a geometric interpretation of the addition law on this Jacobian, obtained by the properties of the linear system of cubics on that curve. We show that our model has mild singularities and that it admits a finite ramified covering to the four-dime

Grzegorz Malczyk, Mihir Kulkarni, Kostas Alexis

This paper introduces a novel semantics-aware inspection planning policy derived through deep reinforcement learning. Reflecting the fact that within autonomous informative path planning missions in unknown environments, it is often only a sparse set of objects of interest that need to be inspected, the method contributes an end-to-end policy that simultaneo

Mete Ismayilzada, Antonio Laverghetta, Simone A. Luchini, Reet Patel

While Large Language Models (LLMs) have demonstrated impressive performance across natural language generation tasks, their ability to generate truly creative content-characterized by novelty, diversity, surprise, and quality-remains limited. Existing methods for enhancing LLM creativity often focus narrowly on diversity or specific tasks, failing to address

Griffen Adams, Ovidiu Costin, Gerald V. Dunne, Sergei Gukov

We show that the fundamental property of preservation of relations, underlying resurgent analysis, provides a new perspective on crossing a natural boundary, an important general problem in theoretical and mathematical physics. This reveals a deeper rigidity of resurgence in a quantum field theory. We study the non-perturbative completion of complex Chern-Si

Filippo Badalamenti, Sampath Kumar Mulagaleti, Mario Eduardo Villanueva, Boris Houska

Configuration-Constrained Tube Model Predictive Control (CCTMPC) offers flexibility by using a polytopic parameterization of invariant sets and the optimization of an associated vertex control law. This flexibility, however, often demands computational trade-offs between set parameterization accuracy and optimization complexity. This paper proposes two innov

Logan A. Pearce, Jared R. Males, Sebastiaan Y. Haffert, Laird M. Close

Most known white dwarfs in multiple systems with main sequence stars have been discovered with M-type companions, because the white dwarf causes detectable UV excess and bluer colors than expected from a single M star. Surveys have shown that the number of white dwarfs in Sirius-like systems within 100 pc of the Sun is lower than expected, suggesting that wh

Yuanbo Fang, Haoze Sun, Jun Liu, Tao Zhang

End-to-end speech large language models ((LLMs)) extend the capabilities of text-based models to directly process and generate audio tokens. However, this often leads to a decline in reasoning and generation performance compared to text input, a phenomenon referred to as intelligence degradation. To systematically evaluate this gap, we propose S2SBench, a be

Dingding Wang, Jianting He, Yizheng Yang, Lei Wu

The emergence of smart contracts brings security risks, exposing users to the threat of losing valuable cryptocurrencies, underscoring the urgency of meticulous scrutiny. Nevertheless, the static analysis of smart contracts in EVM bytecode faces obstacles due to flawed primitives resulting from code reuse introduced by compilers. Code reuse, a phenomenon whe

Yuqiao Tan, Shizhu He, Kang Liu, Jun Zhao

Large Language Models (LLMs) offer a transparent brain with accessible parameters that encode extensive knowledge, which can be analyzed, located and transferred. Consequently, a key research challenge is to transcend traditional knowledge transfer paradigms rooted in symbolic language and achieve genuine Parametric Knowledge Transfer (PKT). Significantly, e

Annika Bush, Meltem Aksoy, Markus Pauly, Greta Ontrup

As organizations increasingly rely on AI systems for decision support in sustainability contexts, it becomes critical to understand the inherent biases and perspectives embedded in Large Language Models (LLMs). This study systematically investigates how five state-of-the-art LLMs -- Claude, DeepSeek, GPT, LLaMA, and Mistral - conceptualize sustainability and

Ange Benise Niyikiza, Zeyu Xiang, Fanghao Zhang, Fengjiao Pan

Boron arsenide (BAs) single crystals had been previously reported to have thermal conductivity of 1500 W/mK at room temperature. Now we achieved thermal conductivity above 2100 W/mK at room temperature in BAs crystals due to much lower concentration of impurities Si, C, and O grown from purified arsenic. We also observed a T-1.8 dependence of the thermal con

Runwu Shi, Zirui Lin, Benjamin Yen, Jiang Wang

This paper aims to achieve single-channel target speech extraction (TSE) in enclosures utilizing distance clues and room information. Recent works have verified the feasibility of distance clues for the TSE task, which can imply the sound source's direct-to-reverberation ratio (DRR) and thus can be utilized for speech separation and TSE systems. However, suc

Eugene Yang, Andrew Yates, Kathryn Ricci, Orion Weller

Retrieve-and-rerank is a popular retrieval pipeline because of its ability to make slow but effective rerankers efficient enough at query time by reducing the number of comparisons. Recent works in neural rerankers take advantage of large language models for their capability in reasoning between queries and passages and have achieved state-of-the-art retriev

F. Golgeleyen, O. Y. Imanuvilov, M. Yamamoto

We consider an elliptic differential inequality: $\vert \Delta u(x) \vert \le C_0(\YYYY^{-\gamma}\vert u(x)\vert + \YYYY^{-\theta}\vert \nabla u(x)\vert)$ in an exterior domain $\R^n \setminus \ooo{U}$, where $U$ is a simply connected bounded domain $U$, $x := (y,z) \in \R^n$ with $y \in \R^m$ and $z\in \R^{n-m}$ for given $m\in \{ 1, ..., n\}$, and $\gamma,

Yu Gao, Yongcun Song, Zhiyu Tan, Hangrui Yue

Elliptic variational inequalities (EVIs) present significant challenges in numerical computation due to their inherent non-smoothness, nonlinearity, and inequality formulations. Traditional mesh-based methods often struggle with complex geometries and high computational costs, while existing deep learning approaches lack generality for diverse EVIs. To allev

Jonas Arruda, Vikas Pandey, Catherine Sherry, Margarida Barroso

Amortized Bayesian inference (ABI) with neural networks has emerged as a powerful simulation-based approach for estimating complex mechanistic models. However, extending ABI to hierarchical models, a cornerstone of modern Bayesian analysis, has been a major hurdle due to the need to simulate and process massive datasets. Our study tackles these challenges by

Riccardo D'Elia

The objective of this proposal is to bridge the gap between Deep Learning (DL) and System Dynamics (SD) by developing an interpretable neural system dynamics framework. While DL excels at learning complex models and making accurate predictions, it lacks interpretability and causal reliability. Traditional SD approaches, on the other hand, provide transparenc

Thomas Sandholm, Sayandev Mukherjee, Lin Cheng, Bernardo A. Huberman

We expand the scope of cache memory to include LEO constellations, which are highly distributed systems with thousands of satellites connected with free-space optics inter-satellite links (ISL) always only one hop from any point on earth. We show how to increase the number of cache hits and improve the speed of inference for the important use case of LLMs. T

Chalamalasetti Kranti, Sherzod Hakimov, David Schlangen

Instruction-tuned large language models (LLMs) have shown strong performance on a variety of tasks; however, generalizing from synthetic to human-authored instructions in grounded environments remains a challenge for them. In this work, we study generalization challenges in spatial grounding tasks where models interpret and translate instructions for buildin

Levin Hornischer, Hannes Leitgeb

We propose a new interpretability method for neural networks, which is based on a novel mathematico-philosophical theory of reasons. Our method computes a vector for each neuron, called its reasons vector. We then can compute how strongly this reasons vector speaks for various propositions, e.g., the proposition that the input image depicts digit 2 or that t

Ona de Gibert, Joseph Attieh, Teemu Vahtola, Mikko Aulamo

We investigate the potential of LLM-generated synthetic data for improving low-resource Machine Translation (MT). Focusing on seven diverse target languages, we construct a document-level synthetic corpus from English Europarl, and extend it via pivoting to 147 additional language pairs. Automatic and human evaluation confirm its overall high quality. We stu

Xutao Mao, Ezra Xuanru Tao, Leyao Wang

Large Language Models (LLMs) are increasingly used as scalable tools for pilot testing, predicting public opinion distributions before deploying costly surveys. To serve as effective pilot testing tools, the performance of these LLMs is typically benchmarked against their ability to reproduce the outcomes of past structured surveys. This evaluation paradigm,

Zuogong Yue, Xinyi Wang, Victor Solo

Clustering of time series based on their underlying dynamics is keeping attracting researchers due to its impacts on assisting complex system modelling. Most current time series clustering methods handle only scalar time series, treat them as white noise, or rely on domain knowledge for high-quality feature construction, where the autocorrelation pattern/fea

Huopu Zhang, Yanguang Liu, Miao Zhang, Zirui He

Predicting earnings surprises from financial documents, such as earnings conference calls, regulatory filings, and financial news, has become increasingly important in financial economics. However, these financial documents present significant analytical challenges, typically containing over 5,000 words with substantial redundancy and industry-specific termi

Huimin Xu, Xin Mao, Feng-Lin Li, Xiaobao Wu

Process Reward Models (PRMs) have demonstrated promising results in mathematical reasoning, but existing process annotation approaches, whether through human annotations or Monte Carlo simulations, remain computationally expensive. In this paper, we introduce Step COmpression for Process Estimation (SCOPE), a novel compression-based approach that significant

Xue Wang, Tian Zhou, Jinyang Gao, Bolin Ding

We present a joint forecasting framework for time series prediction that contrasts with traditional direct or recursive methods. This framework achieves state-of-the-art performance for our designed foundation model, YingLong, and reveals a novel scaling effect: longer outputs significantly enhance model accuracy due to delayed chain-of-thought reasoning in

Ziyang Yu, Wenbing Huang, Yang Liu

Molecular Dynamics (MD) simulations are essential for understanding the atomic-level behavior of molecular systems, giving insights into their transitions and interactions. However, classical MD techniques are limited by the trade-off between accuracy and efficiency, while recent deep learning-based improvements have mostly focused on single-domain molecules

Menglin Yang, Yifei Zhang, Jialin Chen, Melanie Weber

In the era of foundation models and Large Language Models (LLMs), Euclidean space is the de facto geometric setting of our machine learning architectures. However, recent literature has demonstrated that this choice comes with fundamental limitations. To that end, non-Euclidean learning is quickly gaining traction, particularly in web-related applications wh

John M. Davis, Amador Garcia-Fuente, Jaime Ferrer, Salvador Barraza-Lopez

A reference lattice, away from which elastic distortions induced by the spin texturing of 2D magnets take hold, is motivated from a picture of pairwise Biot-Savart interactions among identical solenoids that either elongate or compress a (``zero-current'') spring lattice. Applied to a paradigmatic CrSiTe$_3$ monolayer (ML), the reference is given by the aver

Myung Jun Kim, Félix Lefebvre, Gaëtan Brison, Alexandre Perez-Lebel

Table foundation models bring high hopes to data science: pre-trained on tabular data to embark knowledge or priors, they should facilitate downstream tasks on tables. One specific challenge is that of data semantics: numerical entries take their meaning from context, e.g., column name. Pre-trained neural networks that jointly model column names and table en

Chengtang Yao, Lidong Yu, Zhidan Liu, Jiaxi Zeng

The matching formulation makes it naturally hard for the stereo matching to handle ill-posed regions like occlusions and non-Lambertian surfaces. Fusing monocular priors has been proven helpful for ill-posed matching, but the biased monocular prior learned from small stereo datasets constrains the generalization. Recently, stereo matching has progressed by l

Yutong He, Christoph H. Keitel, Matteo Tamburini

The observed millisecond-scale duration is an essential yet mysterious feature of fast radio bursts (FRBs). In this Letter, we link the observed soft gamma-ray counterpart of FRB 200428 to electron-positron pair cascades driven by Compton scattering and the Breit-Wheeler process. We demonstrate that such pair cascades can truncate FRBs to durations down to m

Paweł Batorski, Adrian Kosmala, Paul Swoboda

Effective prompt engineering remains a central challenge in fully harnessing the capabilities of LLMs. While well-designed prompts can dramatically enhance performance, crafting them typically demands expert intuition and a nuanced understanding of the task. Moreover, the most impactful prompts often hinge on subtle semantic cues, ones that may elude human p

Jinzuomu Zhong, Suyuan Liu, Dan Wells, Korin Richmond

Despite growing interest in generating high-fidelity accents, evaluating accent similarity in speech synthesis has been underexplored. We aim to enhance both subjective and objective evaluation methods for accent similarity. Subjectively, we refine the XAB listening test by adding components that achieve higher statistical significance with fewer listeners a

Tullio Ceccherini-Silberstein, Michel Coornaert

Given a dynamical system $(X,f)$ consisting of a compact metrizable space $X$ and a homeomorphism $f \colon X \to X$, an endomorphism of $(X,f)$ is a continuous map of $X$ into itself which commutes with $f$. One says that a dynamical system $(X,f)$ is surjunctive if every injective endomorphism of $(X,f)$ is surjective. An endomorphism of $(X,f)$ is called

Linfeng Yang, Peilun Li, Jinbao Jian

Unit commitment problem (UCP) is a critical component of power market decision-making. However, its computational complexity necessitates effi-cient solution methods. In this work we propose a framework to accelerate the solving process of the UCP, and the data collecting process for two dis-tinct graph neural network (GNN) policy. We at first train a Neural

Aniket Salvi, Gereon Weiss, Mario Trapp

Autonomous systems that rely on Machine Learning (ML) utilize online fault tolerance mechanisms, such as runtime monitors, to detect ML prediction errors and maintain safety during operation. However, the lack of human-interpretable explanations for these errors can hinder the creation of strong assurances about the system's safety and reliability. This pape

Haoming Huang, Yibo Yan, Jiahao Huo, Xin Zou

Large Language Models (LLMs), despite their remarkable capabilities, are hampered by hallucinations. A particularly challenging variant, knowledge overshadowing, occurs when one piece of activated knowledge inadvertently masks another relevant piece, leading to erroneous outputs even with high-quality training data. Current understanding of overshadowing is

Jiafeng Liang, Shixin Jiang, Xuan Dong, Ning Wang

Large Multimodal Models (LMMs) have recently demonstrated impressive performance on general video comprehension benchmarks. Nevertheless, for broader applications, the robustness of their temporal analysis capability needs to be thoroughly investigated yet predominantly ignored. Motivated by this, we propose a novel temporal robustness benchmark (TemRobBench

Xuecheng Wu, Jiaxing Liu, Danlei Huang, Yifan Wang

Visual-Interleaved Chain-of-Thought (VI-CoT) enables Multi-modal Large Language Models (MLLMs) to continually update their understanding and decision space based on step-wise intermediate visual states (IVS), much like a human would, which has demonstrated impressive success in various tasks, thereby leading to emerged advancements in related downstream benc

Zhaohui Yang, Yuxiao Ye, Shilei Jiang, Chen Hu

Recent advances in reasoning language models have witnessed a paradigm shift from short to long CoT pattern. Given the substantial computational cost of rollouts in long CoT models, maximizing the utility of fixed training datasets becomes crucial. Our analysis reveals that negative responses contain valuable components such as self-reflection and error-corr

Mengzhu Wang, Jiao Li, Shanshan Wang, Long Lan

Semi-supervised learning (SSL) has achieved significant progress in medical image segmentation (SSMIS) through effective utilization of limited labeled data. While current SSL methods for medical images predominantly rely on consistency regularization and pseudo-labeling, they often overlook transferable semantic relationships across different clinical domai

Heng Yang, Jack Cole, Yuan Li, Renzhi Chen

The code of nature, embedded in DNA and RNA genomes since the origin of life, holds immense potential to impact both humans and ecosystems through genome modeling. Genomic Foundation Models (GFMs) have emerged as a transformative approach to decoding the genome. As GFMs scale up and reshape the landscape of AI-driven genomics, the field faces an urgent need

H. Witała, J. Golak, R. Skibiński

We investigate polarization states of the outgoing neutron-proton ($np$) pair in elastic polarized neutron and proton scattering, aiming to find unambiguous evidence for entanglement of their spin states. To obtain complete information about these states, we calculate, using the high precision nucleon-nucleon potential AV18, the final polarizations of the ne

Zhujun Jiang, Xiaolin Luo, Wenying Du, Zhiwei Min

Large-scale structure (LSS) analysis in galaxy surveys is a powerful cosmological probe but is limited by tracer bias, which can obscure underlying information and weaken parameter constraints. Existing methods either model bias or restrict analyses to low-density regions, yet their sensitivity to bias remains poorly understood. We propose a novel method bas

Travis Whyte, Andreas Stathopoulos, Eloy Romero

A modification to the setup algorithm for the multigrid preconditioner of Wilson fermions in lattice QCD is presented. A larger basis of test vectors than that used in conventional multigrid is calculated by the smoother and truncated by singular value decomposition on the chiral components of the test vectors. The truncated basis is used to form the prolong

Peter Baile Chen, Yi Zhang, Dan Roth, Samuel Madden

While humans naturally learn and adapt from past experiences, large language models (LLMs) and their agentic counterparts struggle to retain reasoning from previous tasks and apply them in future contexts. To address this limitation, we propose a novel framework, log-augmented generation (LAG) that directly reuses prior computation and reasoning from past lo

Niko Jokela, Jani Kastikainen, Carlos Nunez, José Manuel Penín

Entanglement entropy has proven to be a powerful tool for probing renormalization group (RG) flows in quantum field theories, with c-functions derived from it serving as candidate measures of the effective number of degrees of freedom. While the monotonicity of such c-functions is well established in many settings, notable exceptions occur in theories with a

Gaël Gendron, Jože M. Rožanec, Michael Witbrock, Gillian Dobbie

Causal world models are systems that can answer counterfactual questions about an environment of interest, i.e. predict how it would have evolved if an arbitrary subset of events had been realized differently. It requires understanding the underlying causes behind chains of events and conducting causal inference for arbitrary unseen distributions. So far, th

Seyoung Song, Seogyeong Jeong, Eunsu Kim, Jiho Jin

Evaluating text generation capabilities of large language models (LLMs) is challenging, particularly for low-resource languages where methods for direct assessment are scarce. We propose MUG-Eval, a novel framework that evaluates LLMs' multilingual generation capabilities by transforming existing benchmarks into conversational tasks and measuring the LLMs' a

Mihir Athale, Vishal Vaddina

Recent advancements in Large Language Models (LLMs) have transformed code generation from natural language queries. However, despite their extensive knowledge and ability to produce high-quality code, LLMs often struggle with contextual accuracy, particularly in evolving codebases. Current code search and retrieval methods frequently lack robustness in both

Nadir Durrani, Basel Mousi, Fahim Dalvi

While Knowledge Editing has been extensively studied in monolingual settings, it remains underexplored in multilingual contexts. This survey systematizes recent research on Multilingual Knowledge Editing (MKE), a growing subdomain of model editing focused on ensuring factual edits generalize reliably across languages. We present a comprehensive taxonomy of M

John M. Campbell

Each of Ramanujan's series for $\frac{1}{\pi}$ is of the form $$ \sum_{n=0}^{\infty} z^n \frac{ (a_{1})_{n} (a_{2})_{n} (a_{3})_{n} }{ (b_{1})_{n} (b_{2})_{n} (b_{3})_{n} } (c_{1} n + c_2) $$ for rational parameters such that the difference between the arguments of any lower and upper Pochhammer symbols is not an integer. In accordance with the work of Chu,

Zhaohui Yang, Chenghua He, Xiaowen Shi, Linjing Li

Many studies focus on data annotation techniques for training effective PRMs. However, current methods encounter a significant issue when applied to long CoT reasoning processes: they tend to focus solely on the first incorrect step and all preceding steps, assuming that all subsequent steps are incorrect. These methods overlook the unique self-correction an