May 2025 arXiv papers — page 65

Zihao Liu, Xiaoyu Wu, Wenna Li, Linlin Yang

Video Anomaly Detection (VAD), which aims to detect anomalies that deviate from expectation, has attracted increasing attention in recent years. Existing advancements in VAD primarily focus on model architectures and training strategies, while devoting insufficient attention to evaluation metrics and benchmarks. In this paper, we rethink VAD evaluation metho

João Henrique Andrade, Tao Feng, Paolo Piccione, Minbo Yang

We investigate the qualitative properties of a critical Hartree equation defined on punctured domains. Our study has two main objectives: analyzing the asymptotic behavior near isolated singularities and establishing radial symmetry of positive singular solutions. First, employing asymptotic analysis, we characterize the local behavior of solutions near the

Jiawei Xue, Zhen Yang, Haitao Lin, Ziji Zhang

Graph Contrastive Learning (GCL), which fuses graph neural networks with contrastive learning, has evolved as a pivotal tool in user-item recommendations. While promising, existing GCL methods often lack explicit modeling of hierarchical item structures, which represent item similarities across varying resolutions. Such hierarchical item structures are ubiqu

Daniel Barzilai, Yuval Margalit, Eitan Gronich, Gilad Yehudai

Over-parameterized models have raised concerns about their potential to memorize training data, even when achieving strong generalization. The privacy implications of such memorization are generally unclear, particularly in scenarios where only model outputs are accessible. We study this question in the context of kernel methods, and demonstrate both empiric

Md. Mithun Hossain, Md. Shakil Hossain, Sudipto Chaki, Md. Rajib Hossain

Aspect-Based Sentiment Analysis (ABSA) is a fundamental task in natural language processing, offering fine-grained insights into opinions expressed in text. While existing research has largely focused on resource-rich languages like English which leveraging large annotated datasets, pre-trained models, and language-specific tools. These resources are often u

Yaxuan Li, Yichen Zhu, Junjie Wen, Chaomin Shen

The field of robotics has made significant strides toward developing generalist robot manipulation policies. However, evaluating these policies in real-world scenarios remains time-consuming and challenging, particularly as the number of tasks scales and environmental conditions change. In this work, we demonstrate that world models can serve as a scalable,

Pavel Exner, Andrii Khrabustovskyi

We analyze an approximation of a Laplacian subject to non-local interface conditions of a $\delta'$-type by Neumann Laplacians on a family of Riemannian manifolds with a sieve-like structure. We establish a (kind of) resolvent convergence for such operators, which in turn implies the convergence of spectra and eigenspaces, and demonstrate convergence of the

Jingping Liu, Ziyan Liu, Zhedong Cen, Yan Zhou

Spatial relation reasoning is a crucial task for multimodal large language models (MLLMs) to understand the objective world. However, current benchmarks have issues like relying on bounding boxes, ignoring perspective substitutions, or allowing questions to be answered using only the model's prior knowledge without image understanding. To address these issue

Haitao Lin, Odin Zhang, Jia Xu, Yunfan Liu

The affinity and specificity of protein-molecule binding directly impact functional outcomes, uncovering the mechanisms underlying biological regulation and signal transduction. Most deep-learning-based prediction approaches focus on structures of atoms or fragments. However, quantum chemical properties, such as electronic structures, are the key to unveilin

Kiljae Lee, Ziqi Liu, Weijing Tang, Yuan Zhang

Data Shapley is an important tool for data valuation, which quantifies the contribution of individual data points to machine learning models. In practice, group-level data valuation is desirable when data providers contribute data in batch. However, we identify that existing group-level extensions of Data Shapley are vulnerable to shell company attacks, wher

Xianlong Zeng, Jun Fang, Peilan Wang, Weidong Mei

Movable antennas (MAs) have emerged as a disruptive technology in wireless communications for enhancing spatial degrees of freedom through continuous antenna repositioning within predefined regions, thereby creating favorable channel propagation conditions. In this paper, we study the problem of position optimization for MA-enabled multi-user MISO systems, w

Yu-Te Hsu, Yidi Liu, Yoshimitsu Kohama, Tommy Kotte

Unconventional superconductivity typically emerges out of a strongly correlated normal state, manifesting as a highly renormalized Fermi liquid or a strange metal with $T$-linear resistivity. In Ruddlesden-Popper bilayer nickelates, superconductivity with a critical temperature $T_{\rm c}$ exceeding 80 and 40~K has been respectively realised in pressurized b

Md. Mithun Hossain, Md. Shakil Hossain, Sudipto Chaki, M. F. Mridha

Multi-modal learning has emerged as a crucial research direction, as integrating textual and visual information can substantially enhance performance in tasks such as classification, retrieval, and scene understanding. Despite advances with large pre-trained models, existing approaches often suffer from insufficient cross-modal interactions and rigid fusion

Yuhao Sun, Zhiyuan Ma, Xinke Shen, Jinhao Li

Electrophysiological brain signals, such as electroencephalography (EEG), exhibit both periodic and aperiodic components, with the latter often modeled as 1/f noise and considered critical to cognitive and neurological processes. Although various theoretical frameworks have been proposed to account for aperiodic activity, its scale-invariant and long-range t

Andrzej Weber

The article presents four reasons why the elliptic genus is the most general characteristic class that admits a generalization to singular spaces. We prove that the elliptic characteristic class (with an additional factor) is essentially the only characteristic class invariant under certain modifications, such as the Atiyah flop, Grassmannian flops, and modi

Yunan Zeng

We prove the off-diagonal estimates of the bilinear iterated commutators in the two-weight setting. The upper bound is established via sparse domination, and the lower bound is proved by the median method. Our methods are so flexible so that it can be easily extended to the multilinear scenario.

Emily Priyadarshini, Massimo Bartoletti

Decentralized applications are often composed of multiple interconnected smart contracts. This is especially evident in DeFi, where protocols are heavily intertwined and rely on a variety of basic building blocks such as tokens, decentralized exchanges and lending protocols. A crucial security challenge in this setting arises when adversaries target individu

Daniel Csizmadia, Andrei Codreanu, Victor Sim, Vighnesh Prabhu

We present Distill CLIP (DCLIP), a fine-tuned variant of the CLIP model that enhances multimodal image-text retrieval while preserving the original model's strong zero-shot classification capabilities. CLIP models are typically constrained by fixed image resolutions and limited context, which can hinder their effectiveness in retrieval tasks that require fin

Ryotaro Arakawa, Sachio Komori, Tomoyasu Taniyama

Magnetic properties of La$_{1-x}$Sr$_x$MnO$_{3}$ (LSMO) are highly sensitive to various factors such as the Sr doping level $x$, lattice strain, and oxygen stoichiometry due to the strongly correlated nature of $3d$ electrons. For the development of energy-efficient spintronic devices with ultra-low magnetic damping of LSMO, a thorough understanding of its c

Hyunwoo Kim, Jaeseong Lee, Sunpyo Hong, Changmin Han

In-host shared memory (IVSHMEM) enables high-throughput, zero-copy communication between virtual machines, but today's implementations lack any security control, allowing any application to eavesdrop or tamper with the IVSHMEM region. This paper presents Secure IVSHMEM, a protocol that provides end-to-end mutual authentication and fine-grained access enforce

Tianming Liu, Manzi Li, Yafeng Yin

The advent of large language models (LLMs) presents new opportunities for travel demand modeling. However, behavioral misalignment between LLMs and humans presents obstacles for the usage of LLMs, and existing alignment methods are frequently inefficient or impractical given the constraints of typical travel demand data. This paper introduces a novel framewo

Jin Zhu, Xin Zhou, Jiaang Yao, Gholamali Aminian

Offline reinforcement learning (RL) aims to learn an optimal policy from pre-collected data. However, it faces challenges of distributional shift, where the learned policy may encounter unseen scenarios not covered in the offline data. Additionally, numerous applications suffer from a scarcity of labeled reward data. Relying on labeled data alone often leads

Manos Chatzakis, Yannis Papakonstantinou, Themis Palpanas

Approximate Nearest Neighbor Search (ANNS) presents an inherent tradeoff between performance and recall (i.e., result quality). Each ANNS algorithm provides its own algorithm-dependent parameters to allow applications to influence the recall/performance tradeoff of their searches. This situation is doubly problematic. First, the application developers have t

Yunxin Li, Xinyu Chen, Zitao Li, Zhenyu Liu

Applying Reinforcement Learning (RL) to Video Large Language Models (Video-LLMs) shows significant promise for complex video reasoning. However, popular Reinforcement Fine-Tuning (RFT) methods, such as outcome-based Group Relative Policy Optimization (GRPO), are limited by data preparation bottlenecks (e.g., noise or high cost) and exhibit unstable improveme

Xurong Liang, Tong Chen, Wei Yuan, Hongzhi Yin

As recommendation services scale rapidly and their deployment now commonly involves resource-constrained edge devices, GNN-based recommender systems face significant challenges, including high embedding storage costs and runtime latency from graph propagations. Our previous work, LEGCF, effectively reduced embedding storage costs but struggled to maintain re

Andrew Luka, Yakir Vizel

Property Directed Reachability (\textsc{Pdr}), also known as IC3, is a state-of-the-art model checking algorithm widely used for verifying safety properties. While \textsc{Pdr} is effective in finding inductive invariants, its underlying proof system, Resolution, limits its ability to construct short proofs for certain verification problems. This paper intro

Reggie C. Pantig, Ali Ovgun, Gaetano Lambiase

We explore the weak-field phenomenology of a compact star spacetime modified by quantum gravitational corrections derived from the effective field theoretical (EFT) approach by Calmet et al. [1]. These corrections, encoded in non-local curvature-squared terms, distinguish matter-supported geometries from vacuum solutions by contributing nontrivial modificati

Bob Junyi Zou, Lu Tian

Hybrid neural ordinary differential equations (neural ODEs) integrate mechanistic models with neural ODEs, offering strong inductive bias and flexibility, and are particularly advantageous in data-scarce healthcare settings. However, excessive latent states and interactions from mechanistic models can lead to training inefficiency and over-fitting, limiting

Carlos Jude G. Maminta, Isaiah Job Enriquez, Deandre Nigel Nunez, Michael B. Dela Fuente

This study presents FiLLM, a Filipino-optimized large language model, designed to enhance natural language processing (NLP) capabilities in the Filipino language. Built upon the SeaLLM-7B 2.5 model, FiLLM leverages Low-Rank Adaptation (LoRA) fine-tuning to optimize memory efficiency while maintaining task-specific performance. The model was trained and evalu

Hewen Xiao, Xiuping Liu, Hang Zhao, Jian Liu

We introduce a novel design of parallel-jaw grippers drawing inspiration from pin-pression toys. The proposed pin-pression gripper features a distinctive mechanism in which each finger integrates a 2D array of pins capable of independent extension and retraction. This unique design allows the gripper to instantaneously customize its finger's shape to conform

Ghalia Kaouane, Jean-François Berret, Yannick Cremillieux, Noël Pinaud

Administration of pulmonary surfactant is crucial for the treatment of respiratory distress syndrome (RDS) in preterm infants. The aim of this study is to evaluate the potential of Curosurf atomization via the Endosurf device, a recently developed spray technology, as a promising approach for surfactant delivery in infants with RDS. A comprehensive analysis

Xiao Xu, Shijie Wang, Haifeng Qin, Zhiqiang Zhao

Tobermorite and Calcium Silicate Hydrate (C-S-H) systems are indispensable cement materials but still lack a satisfactory interatomic potential with both high accuracy and high computational efficiency for better understanding their mechanical performance. Here, we develop a Neuroevolution Machine Learning Potential (NEP) with Ziegler-Biersack-Littmark hybri

Tianchen Deng, Wenhua Wu, Junjie He, Yue Pan

3D Gaussian Splatting has recently shown promising results in dense visual SLAM. However, existing 3DGS-based SLAM methods are all constrained to small-room scenarios and struggle with memory explosion in large-scale scenes and long sequences. To this end, we propose VPGS-SLAM, the first 3DGS-based large-scale RGBD SLAM framework for both indoor and outdoor

Timo Bisswanger, Anne Schmidt, Frank Volmer, Christoph Stampfer

A spin lifetime anisotropy between in-plane and out-of-plane spins in bilayer graphene (BLG) can be achieved by spin-orbit proximity coupling of graphene to transition metal dichalcogenides. This coupling reduces the in-plane spin lifetime due to proximity-induced spin scattering, while the out-of-plane spin lifetime remains largely unaffected. We show that

Catalina Tan, Yipeng Hu, Shaheer U. Saeed

Accurate tumour segmentation is vital for various targeted diagnostic and therapeutic procedures for cancer, e.g., planning biopsies or tumour ablations. Manual delineation is extremely labour-intensive, requiring substantial expert time. Fully-supervised machine learning models aim to automate such localisation tasks, but require a large number of costly an

Varun Jain, Zongwei Wu, Quan Zou, Louis Florentin

This paper presents a comprehensive review of the 1st Challenge on Video Quality Enhancement for Video Conferencing held at the NTIRE workshop at CVPR 2025, and highlights the problem statement, datasets, proposed solutions, and results. The aim of this challenge was to design a Video Quality Enhancement (VQE) model to enhance video quality in video conferen

An Luo, Xun Xian, Jin Du, Fangqiao Tian

Large language models (LLMs) have advanced the automation of data science workflows. Yet it remains unclear whether they can critically leverage external domain knowledge as human data scientists do in practice. To answer this question, we introduce AssistedDS (Assisted Data Science), a benchmark designed to systematically evaluate how LLMs handle domain kno

Zhiwei Lin, Yongtao Wang

Current perception models have achieved remarkable success by leveraging large-scale labeled datasets, but still face challenges in open-world environments with novel objects. To address this limitation, researchers introduce open-set perception models to detect or segment arbitrary test-time user-input categories. However, open-set models rely on human invo

Tianyu Zhang, Xinyu Wang, Lu Li, Zhenghan Tai

While diffusion models have revolutionized text-to-image generation with their ability to synthesize realistic and diverse scenes, they continue to struggle to generate consistent and legible text within images. This shortcoming is commonly attributed to the locality bias inherent in diffusion-based generation, which limits their ability to model long-range

Ibuki Kuroyanagi, Tatsuya Komatsu

We propose a self-supervised learning method using multiple sampling strategies to obtain general-purpose audio representation. Multiple sampling strategies are used in the proposed method to construct contrastive losses from different perspectives and learn representations based on them. In this study, in addition to the widely used clip-level sampling stra

Jun Zhang, Tong Zhang, Ying Wang

Precise and timely simulation of a structure's dynamic behavior is crucial for evaluating its performance and assessing its health status. Traditional numerical methods are often limited by high computational costs and low efficiency, while deep learning approaches offer a promising alternative. However, these data-driven methods still face challenges, such

Haotian Sun, Yitong Li, Yuchen Zhuang, Niao He

Contrastive Language-Image Pretraining (CLIP) has demonstrated strong zero-shot performance across diverse downstream text-image tasks. Existing CLIP methods typically optimize a contrastive objective using negative samples drawn from each minibatch. To achieve robust representation learning, these methods require extremely large batch sizes and escalate com

Ibuki Kuroyanagi, Tomoki Hayashi, Kazuya Takeda, Tomoki Toda

We introduce Serial-OE, a new approach to anomalous sound detection (ASD) that leverages small amounts of anomalous data to improve the performance. Conventional ASD methods rely primarily on the modeling of normal data, due to the cost of collecting anomalous data from various possible types of equipment breakdowns. Our method improves upon existing ASD sys

Huan Wang, Haoran Li, Huaming Chen, Jun Yan

With the advancement of edge computing, federated learning (FL) displays a bright promise as a privacy-preserving collaborative learning paradigm. However, one major challenge for FL is the data heterogeneity issue, which refers to the biased labeling preferences among multiple clients, negatively impacting convergence and model performance. Most previous FL

Ibuki Kuroyanagi, Takuya Fujimura, Kazuya Takeda, Tomoki Toda

This paper addresses performance degradation in anomalous sound detection (ASD) when neither sufficiently similar machine data nor operational state labels are available. We present an integrated pipeline that combines three complementary components derived from prior work and extends them to the unlabeled ASD setting. First, we adapt an anomaly score based

Runze Wang

A strong edge-coloring of a graph $G$ is an edge-coloring such that any two edges of distance at most two receive distinct colors. The minimum number of colors we need in order to give $G$ a strong edge-coloring is called the strong chromatic index of $G$, denoted by $\chi_s'(G)$. The maximum edge weight of $G$ is defined to be $\max\{d(u)+d(v):\ uv\in E(G)\

Miguel Angel Peñaloza Perez, Bruno Lopez Orozco, Jesus Tadeo Cruz Soto, Michelle Bruno Hernandez

Existing mathematical reasoning benchmarks are predominantly English only or translation-based, which can introduce semantic drift and mask languagespecific reasoning errors. To address this, we present AI4Math, a benchmark of 105 original university level math problems natively authored in Spanish. The dataset spans seven advanced domains (Algebra, Calculus

Yong-Gyu Choi, Wansu Kim, Junyeong Park

Given a maximal order $D$ of a central division algebra over a global function field $F$, we prove an explicit sufficient condition for moduli stacks of $D^\times$-shtukas to be proper over a finite field in terms of the local invariants of $D$ and bounds. Our proof is a refinement of E.~Lau's result (Duke Math. J. 140 (2007)), which showed the properness of

Pingbang Hu, Joseph Melkonian, Weijing Tang, Han Zhao

Gradient-based data attribution methods, such as influence functions, are critical for understanding the impact of individual training samples without requiring repeated model retraining. However, their scalability is often limited by the high computational and memory costs associated with per-sample gradient computation. In this work, we propose GraSS, a no

Aotao Wang, Haikuo Shao, Shaobo Ma, Zhongfeng Wang

State Space Models (SSMs), like recent Mamba2, have achieved remarkable performance and received extensive attention. However, deploying Mamba2 on resource-constrained edge devices encounters many problems: severe outliers within the linear layer challenging the quantization, diverse and irregular element-wise tensor operations, and hardware-unfriendly nonli

Yanping Chen, Xueting Han

In this paper, we establish sparse dominations for the Dunkl-Calder\'on-Zygmund operators and their commutators in the Dunkl setting. As applications, we first define the Dunkl-Muckenhoupt $A_p$ weight and obtain the weighted bounds for the Dunkl-Calder\'on-Zygmund operators, as well as the two-weight bounds for their commutators. Moreover, we also obtain th

Sarang Patil, Ashish Parmanand Pandey, Ioannis Koutis, Mengjia Xu

Selective state-space models excel at long-sequence modeling, but their capacity for language representation -- in complex hierarchical reasoning -- remains underexplored. Most large language models rely on \textit{flat} Euclidean embeddings, limiting their ability to capture latent hierarchies. To address this, we propose {\it Hierarchical Mamba (HiM)}, int

Minsu Kim, Pingchuan Ma, Honglie Chen, Stavros Petridis

This paper explores multi-modal controllable Text-to-Speech Synthesis (TTS) where the voice can be generated from face image, and the characteristics of output speech (e.g., pace, noise level, distance, tone, place) can be controllable with natural text description. Specifically, we aim to mitigate the following three challenges in face-driven TTS systems. 1

Abhijit Chakraborty, Chahana Dahal, Ashutosh Balasubramaniam, Tejas Anvekar

We revisit the efficacy of simple, real-valued embedding models for knowledge graph completion and introduce RelatE, an interpretable and modular method that efficiently integrates dual representations for entities and relations. RelatE employs a real-valued phase-modulus decomposition, leveraging sinusoidal phase alignments to encode relational patterns suc

Bowen Wei, Mehrdad Fazli, Ziwei Zhu

Large language models (LLMs) have demonstrated impressive performance on natural language tasks, but their decision-making processes remain largely opaque. Existing explanation methods either suffer from limited faithfulness to the model's reasoning or produce explanations that humans find difficult to understand. To address these challenges, we propose \tex

Bingyang Wang, Yijiang Li, Yitong Qiao, Maijunxian Wang

Cognitive control, the ability to coordinate competing information sources in pursuit of goals, is fundamental to intelligent behavior. We systematically investigate whether Vision Language Models (VLMs) exhibit cognitive control and how computational resources modulate conflict resolution. We construct a benchmark of 4,410 tasks across seven conflict paradi

B. X. Zheng, T. S. Chan, E. H. van Brummelen, J. H. Snoeijer

The deposition of droplets onto a swollen polymer network induces the formation of a wetting ridge at the contact line. Current models typically consider either viscoelastic effects or poroelastic effects, while polymeric gels often exhibit both properties. In this study, we investigate the growth of the wetting ridge using a comprehensive large deformation

Karen Ardila, Aashka Mohite, Abdoljalil Addeh, Amanda V. Tyndall

Brain aging trajectories differ between males and females, yet the genetic factors underlying these differences remain underexplored. Using structural MRI and genotyping data from 40,940 UK Biobank participants (aged 45-83), we computed Brain Age Gap Estimates (BrainAGE) for total brain, hippocampal, and ventricular volumes. We conducted sex-stratified genom

Sanchit Sinha, Aidong Zhang

Concept-based Models are a class of inherently explainable networks that improve upon standard Deep Neural Networks by providing a rationale behind their predictions using human-understandable `concepts'. With these models being highly successful in critical applications like medical diagnosis and financial risk prediction, there is a natural push toward the

Nuowei Liu, Jiahao Kuang, Yanting Liu, Tao Ji

Protein design is a fundamental challenge in biotechnology, aiming to design novel sequences with specific functions within the vast space of possible proteins. Recent advances in deep generative models have enabled function-based protein design from textual descriptions, yet struggle with structural plausibility. Inspired by classical protein design methods

Andrea Urru, Daniel Seleznev, Yujia Teng, Se Young Park

G-type antiferromagnetic BiFeO$_3$ is shown to be an altermagnet. We present the band structure using an unconventional scheme designed to highlight the distinctive spin splitting which is characteristic of altermagnets. We define and show plots of the spin-splitting function in reciprocal space. We show that the nodal surfaces of the spin-splitting function

Ke-Jung Chen, Meng-Yuan Ho, Pei-Cheng Tung

We present new simulations of the formation and evolution of the first star-forming cloud within a massive minihalo of mass of $1.05 \times 10^7\, M_{\odot}$, carried out using the GIZMO code with detailed modeling of primordial gas cooling and chemistry. Unlike previous studies that simulated the formation of the first stars within a smaller cosmological bo

Jeffrey A. Chan-Santiago, Praveen Tirupattur, Gaurav Kumar Nayak, Gaowen Liu

Dataset distillation has emerged as an effective strategy, significantly reducing training costs and facilitating more efficient model deployment. Recent advances have leveraged generative models to distill datasets by capturing the underlying data distribution. Unfortunately, existing methods require model fine-tuning with distillation losses to encourage d

Xiaoqiang Wang, Suyuchen Wang, Yun Zhu, Bang Liu

Chain-of-thought (CoT) reasoning enables large language models (LLMs) to move beyond fast System-1 responses and engage in deliberative System-2 reasoning. However, this comes at the cost of significant inefficiency due to verbose intermediate output. Recent latent-space reasoning methods improve efficiency by operating on hidden states without decoding into

Joseph Geraci, Bessi Qorri, Christian Cumbaa, Mike Tsay

Artificial intelligence (AI) has evolved into an ecosystem of specialized "species," each with unique strengths. We analyze two: DeepSeek-V3, a 671-billion-parameter Mixture of Experts large language model (LLM) exemplifying scale-driven generality, and NetraAI, a dynamical system-based framework engineered for stability and interpretability on small clinica

Rohit Khoja, Devanshu Gupta, Yanjie Fu, Dan Roth

Querying tables with unstructured data is challenging due to the presence of text (or image), either embedded in the table or in external paragraphs, which traditional SQL struggles to process, especially for tasks requiring semantic reasoning. While Large Language Models (LLMs) excel at understanding context, they face limitations with long input sequences.

Kumar Gaurav, Deepayan Banik, Ishan Sharma

Recent space missions have provided substantial evidence of regolith movement on the surfaces of near Earth asteroids. To investigate this phenomenon, we present a continuum-based model that describes regolith motion on nearly spherical asteroids. The theoretical framework employs a depth averaged approach, traditionally used for simulating terrestrial lands

Samuel B. B. Almeida, J. E. G. Silva, C. A. S. Almeida

We study the influence of a localized Gaussian deformation on massless Dirac fermions confined to a two-dimensional curved surface. Both in-plane and out-of-plane displacements are considered within the framework of elasticity theory. These deformations couple to the Dirac spinors via the spin connection and the vielbeins, leading to a position-dependent Fer

Jiong Wu, Yang Xing, Boxiao Yu, Wei Shao

Most publicly available medical segmentation datasets are only partially labeled, with annotations provided for a subset of anatomical structures. When multiple datasets are combined for training, this incomplete annotation poses challenges, as it limits the model's ability to learn shared anatomical representations among datasets. Furthermore, vision-only f

Alexander Altland, Jeremy van der Heijden, Tobias Micklitz, Moshe Rozali

We investigate the physics of a small group of quantum states defined above the sharply defined ground state of a chaotic ensemble. This `universality class of the first levels' (UFL) is realized in the majority of `synthetic' random matrix models but, for all we know, in only one microscopically defined system: low-dimensional gravity. We discuss th

Yining Pan, Qiongjie Cui, Xulei Yang, Na Zhao

LiDAR-based 3D panoptic segmentation often struggles with the inherent sparsity of data from LiDAR sensors, which makes it challenging to accurately recognize distant or small objects. Recently, a few studies have sought to overcome this challenge by integrating LiDAR inputs with camera images, leveraging the rich and dense texture information provided by th

Yuheng Tang, Hongwei Li, Kaijie Zhu, Michael Yang

Motivated by the success of general-purpose large language models (LLMs) in software patching, recent works started to train specialized patching models. Most works trained one model to handle the end-to-end patching pipeline (including issue localization, patch generation, and patch validation). However, it is hard for a small model to handle all tasks, as

Cenlin Duan, Jianlei Yang, Yikun Wang, Yiou Wang

Processing-in-memory (PIM) is a transformative architectural paradigm designed to overcome the Von Neumann bottleneck. Among PIM architectures, digital SRAM-PIM emerges as a promising solution, offering significant advantages by directly integrating digital logic within the SRAM array. However, rigid crossbar architecture and full array activation pose chall

Divij Chawla, Ashita Bhutada, Do Duc Anh, Abhinav Raghunathan

We assess whether AI systems can credibly evaluate investment risk appetite-a task that must be thoroughly validated before automation. Our analysis was conducted on proprietary systems (GPT, Claude, Gemini) and open-weight models (LLaMA, DeepSeek, Mistral), using carefully curated user profiles that reflect real users with varying attributes such as country

Chen Jia

This work studies knowledge distillation (KD) for large language models (LLMs) through preference optimization. We propose a reward-guided imitation learning framework for sequential KD, formulating a min-max optimization problem between the policy and reward model (RM) to minimize the performance gap between the student and teacher policies. Specifically, t

Saman Sarker Joy, Swakkhar Shatabda

Large-scale multitask benchmarks have driven rapid progress in language modeling, yet most emphasize high-resource languages such as English, leaving Bengali underrepresented. We present BnMMLU, a comprehensive benchmark for measuring massive multitask language understanding in Bengali. BnMMLU spans 41 domains across STEM, humanities, social sciences, and ge

Xinmeng Luan, Gary Scavone

This study investigates the use of an unsupervised, physics-informed deep learning framework to model a one-degree-of-freedom mass-spring system subjected to a nonlinear friction bow force and governed by a set of ordinary differential equations. Specifically, it examines the application of Physics-Informed Neural Networks (PINNs) and Physics-Informed Deep O

Longfei Yun, Chenyang An, Zilong Wang, Letian Peng

Instruction-tuned large language models (LLMs) employ structured templates, such as role markers and special tokens, to enforce format consistency during inference. However, we identify a critical limitation of such formatting: it induces a phenomenon we term diversity collapse, where the model generates semantically similar outputs for open-ended inputs, un

William Merrill, Ashish Sabharwal

Chain of thought is a natural inference-time method for increasing the computational power of transformer-based large language models (LLMs), but comes at the cost of sequential decoding. Are there more efficient alternatives to expand a transformer's expressive power without adding parameters? We consider transformers with padding tokens as a form of parall

Zhenhao Zhang, Ye Shi, Lingxiao Yang, Suting Ni

Understanding and synthesizing realistic 3D hand-object interactions (HOI) is critical for applications ranging from immersive AR/VR to dexterous robotics. Existing methods struggle with generalization, performing well on closed-set objects and predefined tasks but failing to handle unseen objects or open-vocabulary instructions. We introduce OpenHOI, the fi

Yong Xiao, Haoran Zhou, Xubo Li, Yayu Gao

Agentic AI networking (AgentNet) is a novel AI-native networking paradigm that relies on a large number of specialized AI agents to collaborate and coordinate for autonomous decision-making, dynamic environmental adaptation, and complex goal achievement. It has the potential to facilitate real-time network management alongside capabilities for self-configura

Jintao Sun, Hu Zhang, Gangyi Ding, Zhedong Zheng

Modern end-to-end autonomous driving systems suffer from a critical limitation: their planners lack mechanisms to enforce temporal consistency between predicted trajectories and evolving scene dynamics. This absence of self-supervision allows early prediction errors to compound catastrophically over time. We introduce Echo Planning (EchoP), a new self-correc

Muhammad Wahid Akram, Keshav Sood, Muneeb Ul Hassan, Basant Subba

Lately, cybercriminals constantly formulate productive approaches to exploit individuals. This article exemplifies an innovative attack, namely QR-based Browser-in-The-Browser (BiTB), using proficiencies of Large Language Model (LLM) i.e. Google Gemini. The presented attack is a fusion of two emerging attacks: BiTB and Quishing (QR code phishing). Our study

Xuanming Zhang, Yuxuan Chen, Samuel Yeh, Sharon Li

Human social interactions depend on the ability to infer others' unspoken intentions, emotions, and beliefs-a cognitive skill grounded in the psychological concept of Theory of Mind (ToM). While large language models (LLMs) excel in semantic understanding tasks, they struggle with the ambiguity and contextual nuance inherent in human communication. To bridge

Honglin Bao, Siyang Wu, Jiwoong Choi, Yingrong Mao

This paper calls on the research community not only to investigate how human biases are inherited by large language models (LLMs) but also to explore how these biases in LLMs can be leveraged to make society's "unwritten code" - such as implicit stereotypes and heuristics - visible and accessible for critique. We introduce a conceptual framework through a ca

Jaime A. Alvarado-Montes, Mario Sucerquia, Jorge I. Zuluaga, Christian Schwab

TOI-2109b is the ultra-hot Jupiter with the shortest orbital period ($\sim16\,$hr) yet discovered. At this close distance, strong tidal interactions can produce a significant exchange of angular momentum with the star. Since the orbital period of this planet is shorter than the stellar rotation period, TOI-2109b may be an optimal candidate for studying orbit

Sudeep R, Sarojini M, Uma Mahendra Kumar Koppolu

We have investigated the structural stability of a binary compound TiSn in the zincblende symmetry. The phonon dispersion studies confirms that, TiSn with a nominal composition of 1:1 can exist in zincblende form. No imaginary frequencies are observed indicating the stable bonding nature of Ti-Sn. From the First principles calculations based on density funct

Huaqing Cheng, Chen Zhang, Zhixing Ling, Xiaojin Sun

We report on results of the on-ground X-ray calibration of the Wide-field X-ray Telescope (WXT) built from novel lobster-eye micro-pore optics, onboard the Einstein Probe (EP) satellite. To fully characterize the instrumental performance and properties, a series of tests and calibrations have been carried out at different levels of devices, assemblies and th

Michael Clarke, Michael Joffe

The introduction of generative AI tools such as ChatGPT into creative workplaces has sparked highly visible, but binary worker replacement and augmentation debates. This study reframes this argument by examining how creative professionals re-specify a division of labor with these tools. Through 17 ethnomethodologically informed interviews with international

Zahra Bayat, Mark P. Hertzberg

We examine data from the Dark Energy Spectroscopic Instrument (DESI) collaboration which has implications for the nature of dark energy. We consider classes of models that manifestly obey the null energy condition, with a focus on quintessence models. We find that hilltop potentials and exponential potentials provide modest improvement compared to a cosmolog

Ali Lotfian, Ehsan Roohi

Platelet adhesion and aggregation are essential for primary hemostasis, forming a clot that quickly stops initial bleeding. Despite this critical role, the dynamic interactions of platelet receptors with exposed collagen and von Willebrand factor (vWF) at the injury site and how these interactions influence thrombus formation under varying blood flow conditi

Ronny Vallejos, Clemente Ferrer, Jorge Mateu

This paper introduces a novel coefficient for measuring agreement between two lattice sequences observed in the same areal units, motivated by the analysis of different methodologies for measuring poverty rates in Chile. Building on the multivariate concordance coefficient framework, our approach accounts for dependencies in the multivariate lattice process

Dhruv Agarwal, Anya Shukla, Sunayana Sitaram, Aditya Vashistha

Large language models (LLMs) are used worldwide, yet exhibit Western cultural tendencies. Many countries are now building ``regional'' or ``sovereign'' LLMs, but it remains unclear whether they reflect local values and practices or merely speak local languages. Using India as a case study, we evaluate six Indic and six global LLMs on two dimensions -- values

Haitian Zhong, Yuhuan Liu, Ziyang Xu, Guofan Liu

Large language model editing methods frequently suffer from overfitting, wherein factual updates can propagate beyond their intended scope, overemphasizing the edited target even when it's contextually inappropriate. To address this challenge, we introduce REACT (Representation Extraction And Controllable Tuning), a unified two-phase framework designed for p

Hyunho Ha, Lei Xiao, Christian Richardt, Thu Nguyen-Phuoc

We introduce a novel geometry-guided online video view synthesis method with enhanced view and temporal consistency. Traditional approaches achieve high-quality synthesis from dense multi-view camera setups but require significant computational resources. In contrast, selective-input methods reduce this cost but often compromise quality, leading to multi-vie

Ryan Saklad, Aman Chadha, Oleg Pavlov, Raha Moraffah

Understanding and inferring causal relationships from texts is a core aspect of human cognition and is essential for advancing large language models (LLMs) towards artificial general intelligence. Existing work evaluating LLM causal reasoning primarily relies on synthetic or simplified texts with explicitly stated causal relationships. These texts typically

Yanben Shen, Timilehin T. Ayanlade, Venkata Naresh Boddepalli, Mojdeh Saadati

Early weed identification is crucial for effective management and control, and researchers, agronomists, and technology developers are increasingly interested in automating this process using computer vision and artificial intelligence; however, limited expert-verified data and variable morphological features have hindered the development of AI-based weed id

Wenda Zhang

The advancements of Large language models (LLMs) have provided great opportunities to text-to-SQL tasks to overcome the main challenges to understand complex domain information and complex database structures in business applications. In this paper, we propose a meta-aware learning framework to integrate domain knowledge, database schema, chain-of-thought re

Dina Albassam

Unstructured clinical notes contain essential patient information but are challenging for physicians to search and interpret efficiently. Although large language models (LLMs) have shown promise in question answering (QA), most existing systems lack transparency, usability, and alignment with clinical workflows. This work introduces an interactive QA system

Tinglin Huang, Tianyu Liu, Mehrtash Babadi, Wengong Jin

Spatial transcriptomics (ST) has emerged as a powerful technology for bridging histology imaging with gene expression profiling. However, its application has been limited by low throughput and the need for specialized experimental facilities. Prior works sought to predict ST from whole-slide histology images to accelerate this process, but they suffer from t