May 2025 arXiv papers — page 12

Luis Ibanez-Lissen, Lorena Gonzalez-Manzano, Jose Maria de Fuentes, Nicolas Anciaux

Large Language Models (LLMs) are being extensively used for cybersecurity purposes. One of them is the detection of vulnerable codes. For the sake of efficiency and effectiveness, compression and fine-tuning techniques are being developed, respectively. However, they involve spending substantial computational efforts. In this vein, we analyse how Linear Prob

Longjie Luo, Lin Li, Qingyang Hong

Due to the lack of target speech annotations in real-recorded far-field conversational datasets, speech enhancement (SE) models are typically trained on simulated data. However, the trained models often perform poorly in real-world conditions, hindering their application in far-field speech recognition. To address the issue, we (a) propose direct sound estim

Kailin Jiang, Yuntao Du, Yukai Ding, Yuchen Ren

Large Multimodal Models (LMMs) store vast amounts of pretrained knowledge but struggle to remain aligned with real-world updates, making it difficult to avoid capability degradation when acquiring evolving knowledge. Furthermore, most current work focuses on exploring static textual knowledge injection, neglecting dynamic multimodal evolving knowledge inject

Eojin Kang, Jaehyuk Yu, Juae Kim

Recent studies on personas have improved the way Large Language Models (LLMs) interact with users. However, the effect of personas on domain-specific question-answering (QA) tasks remains a subject of debate. This study analyzes whether personas enhance specialized QA performance by introducing two types of persona: Profession-Based Personas (PBPs) (e.g., sc

Kim-Khuong Huynh, Rasmus Baden Stubkjær, Ventrapati Pavankumar, Emilie Skytte Vosegaard

Ideal crystals are fully ordered, but real-world crystals always contain defects breaking translational symmetry. Random defects in crystals have important implications and they e.g. provide the foundation for semiconductor-based electronic devices. Structurally correlated defects introduce an additional level of complexity, which may lead to novel materials

Boris Zilber

The aim of this note is to recast somewhat informal axiom system of quantum mechanics used by physicists (Dirac calculus) in the language of Continuous Logic. We note an analogy between Tarski's notion of cylindric algebras, as a tool of algebraisation of first order logic, and Hilbert spaces which can serve the same purpose for continuous logic of physics.

Longjie Luo, Shenghui Lu, Lin Li, Qingyang Hong

This paper presents our system for the MISP-Meeting Challenge Track 2. The primary difficulty lies in the dataset, which contains strong background noise, reverberation, overlapping speech, and diverse meeting topics. To address these issues, we (a) designed G-SpatialNet, a speech enhancement (SE) model to improve Guided Source Separation (GSS) signals; (b)

Konstantin Aal, Tanja Aal, Vasil Navumau, David Unbehaun

This paper explores the emotional, ethical and practical dimensions of integrating Artificial Intelligence (AI) into personal and professional workflows, focusing on the concept of feeling guilty as a 'c(ai)borg' - a human augmented by AI. Inspired by Donna Haraway's Cyborg Manifesto, the study explores how AI challenges traditional notions of creativity, or

Xin Chen, Yarden As, Andreas Krause

Large language models (LLMs) have emerged as powerful tools but pose significant safety risks through harmful outputs and vulnerability to adversarial attacks. We propose SaP, short for Safety Polytope, a geometric approach to LLM safety that learns and enforces multiple safety constraints directly in the model's representation space. We develop a framework

S. D. Cardell, A. Fúster-Sabater, V. Requena, M. Beltrá

Boolean functions and binary sequences are main tools used in cryptography. In this work, we introduce a new bijection between the set of Boolean functions and the set of binary sequences with period a power of two. We establish a connection between them which allows us to study some properties of Boolean functions through binary sequences and vice versa. Th

Shebha Anandhi Jegadeesan, Yujie Zhao, Graham M. Smith, Ilya Kuprov

In pulsed dynamic nuclear polarization (DNP), enhancement of the polarization of bulk nuclei requires the repeated application of a microwave pulse sequence. So far, analysis of a one-time transfer of electron spin polarization to a dipolar-coupled nuclear spin has guided the design of DNP pulse sequences. This has obvious shortcomings, such as an inability

Heejo Kong, Sung-Jin Kim, Gunho Jung, Seong-Whan Lee

Conventional semi-supervised learning (SSL) ideally assumes that labeled and unlabeled data share an identical class distribution, however in practice, this assumption is easily violated, as unlabeled data often includes unknown class data, i.e., outliers. The outliers are treated as noise, considerably degrading the performance of SSL models. To address thi

Zhentao Xie, Chengcheng Han, Jinxin Shi, Wenjun Cui

Although multi-agent systems based on large language models show strong capabilities on multiple tasks, they are still limited by high computational overhead, information loss, and robustness. Inspired by ResNet's residual learning, we propose Residual Mixture-of-Agents (RMoA), integrating residual connections to optimize efficiency and reliability. To maxim

Chunxu Liu, Chi Xie, Xiaxu Chen, Wei Li

Text-to-Image Retrieval (T2IR) is a highly valuable task that aims to match a given textual query to images in a gallery. Existing benchmarks primarily focus on textual queries describing overall image semantics or foreground salient objects, possibly overlooking inconspicuous small objects, especially in complex environments. Such small object retrieval is

Akaki Mamageishvili, Benny Sudakov

We compare the total capital efficiency of secure restaking and Proof-of-Stake (PoS) protocols. First, we consider the sufficient condition for the restaking graph to be secure. The condition implies that it is always possible to transform such a restaking graph into separate secure PoS protocols. Next, we derive two main results: upper and lower bounds on t

A. Mandziak, V. Sosa, P. Nita, L. Martín-García

We have measured the circular magnetic dichroism in the x-ray absorption at the K-edge of oxygen in microcrystals of different spinel oxides. The microcrystals are islands of micrometric size and nanometric thickness, grown on Ru(0001) substrates using high-temperature oxygen-assisted molecular beam epitaxy. The domains observed in the oxygen K-edge dichrois

Jin Wang, Wenbin Jiang, Xiangbo Wang, Yubo You

Neural audio compression has emerged as a promising technology for efficiently representing speech, music, and general audio. However, existing methods suffer from significant performance degradation at limited bitrates, where the available embedding space is sharply constrained. To address this, we propose a universal high-fidelity neural audio compression

Jorge Castillo-Mateo, Zeus Gracia-Tabuenca, Jesús Asín, Ana C. Cebrián

Record-breaking temperature events are now frequently in the news, proffered as evidence of climate change, and often bring significant economic and human impacts. Our previous work undertook the first substantial spatial modelling investigation of temperature record-breaking across years for any given day within the year, employing a dataset consisting of o

Paweł Magnuszewski, Sylwester Arabas

In this paper, we discuss a simple yet robust PDE method for evaluating path-dependent Asian-style options using the non-oscillatory forward-in-time second-order MPDATA finite-difference scheme. The valuation methodology involves casting the Black-Merton-Scholes equation as a transport problem by first transforming it into a homogeneous advection-diffusion P

Md Shahriar Rahim Siddiqui, Moshe Eliasof, Eldad Haber

Flow matching casts sample generation as learning a continuous-time velocity field that transports noise to data. Existing flow matching networks typically predict each point's velocity independently, considering only its location and time along its flow trajectory, and ignoring neighboring points. However, this pointwise approach may overlook correlations b

Laura Müller, Piklu Mallick, Antonio B. Marín-Carballo, Philipp Dönges

HIV pre-exposure Prophylaxis (PrEP) has become essential for global HIV control, but its implementation coincides with rising bacterial STI rates among men who have sex with men (MSM). While risk-compensation behavioral changes like reduced condom use are frequently reported, we examine whether intensified asymptomatic screening in PrEP programs creates surv

David Sherrington, Scott Kirkpatrick

In 1975, two papers were published that together sparked major new directions, conceptual, mathematical and practically applicable, in several previously disparate fields of science. In this short review, we expose key aspects of their thinking, implementations and implications, along with a selection of further crucial and consequential developments. These

Bozhong Zheng, Jinye Gan, Xiaohao Xu, Xintao Chen

3D point cloud anomaly detection is essential for robust vision systems but is challenged by pose variations and complex geometric anomalies. Existing patch-based methods often suffer from geometric fidelity issues due to discrete voxelization or projection-based representations, limiting fine-grained anomaly localization. We introduce Pose-Aware Signed Dist

Deep H. Makadiya

This thesis investigates certain structural properties of twisted Chevalley groups over commutative rings, focusing on three key problems. Let $R$ be a commutative ring satisfying mild conditions. Let $G_{\pi,\sigma} (\Phi, R)$ denote a twisted Chevalley group over $R$, and let $E'_{\pi, \sigma} (\Phi, R)$ denote its elementary subgroup. The first problem co

Giovanny A. Cuervo-Londoño, Javier Sánchez, Ángel Rodríguez-Santana

Oceanographic forecasting impacts various sectors of society by supporting environmental conservation and economic activities. Based on global circulation models, traditional forecasting methods are computationally expensive and slow, limiting their ability to provide rapid forecasts. Recent advances in deep learning offer faster and more accurate prediction

Benoit Cloitre

A family of nested recurrence relations $a(n+1) = n - a^{(m)}(n) + a^{(m+1)}(n)$, parameterized by an integer $m \ge 1$ with initial condition $a(1)=1$, is studied. We prove that $a(n)=n-h(n)$ is the unique solution satisfying this condition, where $h(n)$ is an arithmetical sequence in which each non-negative integer $k$ appears $mk+1$ times, with $h(n)$ 1-i

Xu Wang, Zihao Li, Benyou Wang, Yan Hu

Large language models (LLMs) store vast amounts of information, making them powerful yet raising privacy and safety concerns when selective knowledge removal is required. Existing unlearning strategies, ranging from gradient-based fine-tuning and model editing to sparse autoencoder (SAE) steering, either lack interpretability or fail to provide a robust defe

Christopher Bagdon, Aidan Combs, Carina Silberer, Roman Klinger

Accurate modeling of subjective phenomena such as emotion expression requires data annotated with authors' intentions. Commonly such data is collected by asking study participants to donate and label genuine content produced in the real world, or create content fitting particular labels during the study. Asking participants to create content is often simpler

David Gamez

Over the last thirty years, considerable progress has been made with the development of systems that can drive cars, play games, predict protein folding and generate natural language. These systems are described as intelligent and there has been a great deal of talk about the rapid increase in artificial intelligence and its potential dangers. However, our t

Mainak Bhowmik, Mihai Putinar

Bounded holomorphic interpolation problems associated to finitely many data have, in general, distinct solutions. Uniqueness arises only in some convex extreme configurations. Rational inner functions in a polydisk are the best understood examples in this sense. We analyze the continuity of global solutions as functions of the finite interpolation data in ne

Amit Peleg, Naman Deep Singh, Matthias Hein

Vision-language models like CLIP have demonstrated remarkable zero-shot capabilities in classification and retrieval. However, these models often struggle with compositional reasoning - the ability to understand the relationships between concepts. A recent benchmark, SugarCrepe++, reveals that previous works on improving compositionality have mainly improved

Zhiwei Liu, Lingfei Qian, Qianqian Xie, Jimin Huang

Large language models and vision-language models (which we jointly call LMs) have transformed NLP and CV, demonstrating remarkable potential across various fields. However, their capabilities in affective analysis (i.e. sentiment analysis and emotion detection) remain underexplored. This gap is largely due to the absence of comprehensive evaluation benchmark

Zhenghua Pan, Yong Wang

In the field of artificial intelligence, understanding, distinguishing, expressing, and computing the negation in knowledge is a fundamental issue in knowledge processing and research. In this paper, we examine and analyze the understanding and characteristics of negation in various fields such as philosophy, logic, and linguistics etc. Based on the distinct

Farzin Safarzadeh-Maleki

We present a single analytic scale factor \(a(t)=e^{H(t)t}\bigl(1-e^{-k(t)t}\bigr)^{b(t)}\) that unifies the cosmic expansion history from inflation to the present in a smooth and differentiable manner. By eliminating conventional piecewise definitions, the model provides a compact global description of cosmic evolution while remaining consistent with SNIa,

Kaiwen Shen, Ping Tang, Xianzhe Chen, Yifan Gao

Ferroelectrics feature spontaneous electric dipolar order reconfigurable via electric fields. Recent theoretical studies of the collective excitations of this electric dipolar order give rise to the hope that "ferron" quasiparticles may complement the magnons of magnetic materials in information and heat management technologies. Yet direct experimental evide

Hiroshi Matano, Shuichi Jimbo

We consider a bistable reaction-diffusion equation on a metric graph that is a generalization of the so-called star graphs. More precisely, our graph $\Omega$ consists of a bounded finite metric graph $D$ of arbitrary configuration and a finite number of branches $\Omega_1,\ldots,\Omega_N\,(N\geq 2)$ of infinite length emanating from some of the vertices of

Runnan Lu, Yuxuan Zhang, Jiaming Liu, Haofan Wang

Generating accurate multilingual text with diffusion models has long been desired but remains challenging. Recent methods have made progress in rendering text in a single language, but rendering arbitrary languages is still an unexplored area. This paper introduces EasyText, a text rendering framework based on DiT (Diffusion Transformer), which connects deno

Byung Hee An, Jang Soo Kim

The $ k $-configuration space $ B_k\Gamma $ of a topological space $ \Gamma $ is the space of sets of $ k $ distinct points in $ \Gamma $. In this paper, we consider the case where $ \Gamma $ is a graph of circumference at most $1$. We show that for all $ k\ge0 $, the $ i $-th Betti number of $ B_k\Gamma $ is given by a polynomial $P_\Gamma^i(k)$ in $ k $, c

V. I. Tokar, H. Dreyssé

The thinning method for numerical generation of the nonhomogeneous Poisson process (NHPP) arrival times has been adapted to accelerate Monte Carlo simulations of the kinetic Ising models (KIMs) with the Glauber spin-flip dynamics. The performance of the suggested algorithms has been illustrated by simulation of the decay of metastable states in stationary KI

Yang Sui, Qi Xu, Yang Bai, Annie Qu

Multi-task learning (MTL) has emerged as an imperative machine learning tool to solve multiple learning tasks simultaneously and has been successfully applied to healthcare, marketing, and biomedical fields. However, in order to borrow information across different tasks effectively, it is essential to utilize both homogeneous and heterogeneous information. A

Agniva Das, Muralidharan K

The Himalayan region, including Nepal, is prone to frequent and large earthquakes. Accurate forecasting of these earthquakes is crucial for minimizing loss of life and damage to infrastructure. In this study, we propose various time-scaled Epidemic Type Aftershock Sequence (ETAS) models to forecast earthquakes in Nepal. The ETAS model is a statistical model

Feng Chen, Kanokphan Lertniphonphan, Qiancheng Yan, Xiaohui Fan

This report introduces our team's (PCIE_EgoPose) solutions for the EgoExo4D Pose and Proficiency Estimation Challenges at CVPR2025. Focused on the intricate task of estimating 21 3D hand joints from RGB egocentric videos, which are complicated by subtle movements and frequent occlusions, we developed the Hand Pose Vision Transformer (HP-ViT+). This architect

Meng Ji

In this paper, we study the regularity of the solution for the obstacle problem associated with the linearized Monge-Amp\`ere operator: \begin{align*} \begin{cases} &u\geq\varphi \text{\quad in } \Omega &L_{ w}u=\tr( W D^{2}u)\leq 0 \text{\quad in } \Omega &L_{ w}u= 0 \text{\quad in } \{u>\varphi\} &u=0 \text{\quad on } \partial\Omega, \end{cases} \end{align

Eojin Kang, Juae Kim

Multilingual large language models (LLMs) offer promising opportunities for cross-lingual information access, yet their use of factual knowledge remains highly sensitive to the input language. Prior work has addressed this through English prompting and evaluation, assuming that English-based reasoning is universally beneficial. In this work, we challenge tha

Samsuzzaman Afroz, Sanjib Kumar Agarwalla, Dipankar Bhattacharya, Soumya Bhattacharya

The multi-messenger science using different observational windows to the Universe such as Gravitational Waves (GWs), Electromagnetic Waves (EMs), Cosmic Rays (CRs), and Neutrinos offer an opportunity to study from the scale of a neutron star to cosmological scales over a large cosmic time. At the smallest scales, we can explore the structure of the neutron s

Wenlong Jiao, Binglong Li, Wei Shang, Ping Wang

Image deblurring plays a crucial role in enhancing visual clarity across various applications. Although most deep learning approaches primarily focus on sRGB images, which inherently lose critical information during the image signal processing pipeline, RAW images, being unprocessed and linear, possess superior restoration potential but remain underexplored.

Hanting Wang, Tao Jin, Wang Lin, Shulei Wang

Bridge models in image restoration construct a diffusion process from degraded to clear images. However, existing methods typically require training a bridge model from scratch for each specific type of degradation, resulting in high computational costs and limited performance. This work aims to efficiently leverage pretrained generative priors within existi

Jiahao Su, Ji Liu, Jianyu Li, Zhangkai Cao

The Cooper pair Bose metal (CPBM) is a non-superfluid quantum phase in which uncondensed fermion pairs form a "Bose surface" in momentum space. We investigate the CPBM in the two-dimensional spin-anisotropic attractive Hubbard model by tuning the next-nearest-neighbor (NNN) hopping t', carrier filling n, and spin anisotropy alpha, using large-scale constrain

Kanokphan Lertniphonphan, Feng Chen, Junda Xu, Fengbu Lan

This report presents our team's PCIE_Interaction solution for the Ego4D Social Interaction Challenge at CVPR 2025, addressing both Looking At Me (LAM) and Talking To Me (TTM) tasks. The challenge requires accurate detection of social interactions between subjects and the camera wearer, with LAM relying exclusively on face crop sequences and TTM combining spe

Giannis Nikolentzos, Konstantinos Skianis

The Lipschitz constant of a neural network is connected to several important properties of the network such as its robustness and generalization. It is thus useful in many settings to estimate the Lipschitz constant of a model. Prior work has focused mainly on estimating the Lipschitz constant of multi-layer perceptrons and convolutional neural networks. Her

Mika Feng, Koichi Ito, Takafumi Aoki, Tetsushi Ohki

Face recognition systems are designed to be robust against changes in head pose, illumination, and blurring during image capture. If a malicious person presents a face photo of the registered user, they may bypass the authentication process illegally. Such spoofing attacks need to be detected before face recognition. In this paper, we propose a spoofing atta

Xianheng Ma, Hongchen Tan, Xiuping Liu, Yi Zhang

In this paper, we leverage the advantages of event cameras to resist harsh lighting conditions, reduce background interference, achieve high time resolution, and protect facial information to study the long-sequence event-based person re-identification (Re-ID) task. To this end, we propose a simple and efficient long-sequence event Re-ID model, namely the Sp

Håkon Tjelmeland, Hanna Bu Kvaløy

Suppose we have a distribution of interest, with density $p(x),x\in {\cal X}$ say, and an algorithm claimed to generate samples from $p(x)$. Moreover, assume we have available a Metropolis--Hastings transition kernel fulfilling detail balance with respect to $p(x)$. In such a situation we formulate a hypothesis test where $H_0$ is that the claimed sampler re

Yifei Cheng, Li Shen, Hao Sun, Nan Yin

Sharpness-Aware Minimization (SAM) optimizer enhances the generalization ability of the machine learning model by exploring the flat minima landscape through weight perturbations. Despite its empirical success, SAM introduces an additional hyper-parameter, the perturbation radius, which causes the sensitivity of SAM to it. Moreover, it has been proved that t

A. Barletta, M. Celli, P. V. Brandão

A statistical analysis of the wall roughness effect is carried out to determine the impact of the shape uncertainty on the Poiseuille number and Nusselt number of laminar forced convection. The focus is on the fully developed regime in a semicircular microchannel where the heat transfer occurs from the diametrical plane boundary, modelled as a perfectly smoo

Soumyakanti Pan, Sudipto Banerjee

Air pollution remains a major environmental risk factor that is often associated with adverse health outcomes. However, quantifying and evaluating its effects on human health is challenging due to the complex nature of exposure data. Recent technological advances have led to the collection of various indicators of air pollution at increasingly high spatial-t

Zhichao Han, Xijie Huang, Zhuxiu Xu, Jiarui Zhang

Quadrotors have demonstrated remarkable versatility, yet their full aerobatic potential remains largely untapped due to inherent underactuation and the complexity of aggressive maneuvers. Traditional approaches, separating trajectory optimization and tracking control, suffer from tracking inaccuracies, computational latency, and sensitivity to initial condit

L. P. Garate-Nuñez, A. S. G. Robotham, S. Bellstedt, L. J. M. Davies

The intra-halo light (IHL) is the diffuse stellar component that surrounds galaxies, groups, and clusters. Its formation is intimately linked to the hierarchical assembly of the system, making it a key tracer of galaxy evolution. However, the low surface brightness (LSB) of the IHL makes it challenging to detect and also to distinguish from the point spread

Vibecke Markhus, Katarina Fritz-Wallace, Olav Mjaavatten, Einar K. Kristoffersen

Background: Platelet proteomics offers valuable insights for clinical research, yet isolating high-purity platelets remains a challenge. Current methods often lead to contamination or platelet loss, compromising data quality and reproducibility. Objectives: This study aimed to optimize a platelet isolation technique that yields high-purity samples with minim

Suhyeon Lee, Yeongju Bak

Optimistic Rollups (ORUs) significantly enhance blockchain scalability but inherently suffer from the verifier's dilemma, particularly concerning validator attentiveness. Current systems lack mechanisms to proactively ensure validators are diligently monitoring L2 state transitions, creating a vulnerability where fraudulent states could be finalized. This pa

Christof Wetterich

The quantum or quantum field theory concept of a complex wave function is useful for understanding the information transport in classical statistical generalized Ising models. We relate complex conjugation to the discrete transformations charge conjugation ($C$), parity ($P$) and time reversal ($T$). A subclass of generalized Ising models are probabilistic c

Kensuke Kakimoto, Shun Uchino

Motivated by recent advances in ultracold atomic gas experiments, we investigate a two-terminal mesoscopic system in which two-body loss occurs locally at the center of a one-dimensional chain. By means of the self-consistent Born approximation in the Keldysh formalism, we uncover mesoscopic current formulas that are experimentally relevant and applicable to

Yuqi Fan, Zhiyong Cui, Zhenning Li, Yilong Ren

Reliable planning is crucial for achieving autonomous driving. Rule-based planners are efficient but lack generalization, while learning-based planners excel in generalization yet have limitations in real-time performance and interpretability. In long-tail scenarios, these challenges make planning particularly difficult. To leverage the strengths of both rul

Liangyang Ouyang, Yuki Sakai, Ryosuke Furuta, Hisataka Nozawa

This paper addresses the task of assessing PICU team's leadership skills by developing an automated analysis framework based on egocentric vision. We identify key behavioral cues, including fixation object, eye contact, and conversation patterns, as essential indicators of leadership assessment. In order to capture these multimodal signals, we employ Aria Gl

Hao Chen, Yukun Yan, Sen Mei, Wanxiang Che

Retrieval-Augmented Generation (RAG) augments Large Language Models (LLMs) with external knowledge to improve factuality. However, existing RAG systems frequently underutilize the retrieved documents, failing to extract and integrate the key clues needed to support faithful and interpretable reasoning, especially in cases where relevant evidence is implicit,

Angela Pistoia, Giuseppe Mario Rago, Giusi Vaira

The paper addresses the existence of multi-bubble solutions for the well-known Brezis-Nirenberg problem. Although there is extensive literature on the subject, the existence of solutions that blow up at multiple points in a 4D bounded domain remains an open problem. The goal of the present paper is to resolve this longstanding issue. In particular, we exhibi

Jared Miller

This work approaches the problem of computing incremental $\ell_1$ and $\ell_\infty$ gains for discrete-time positive systems in \lure feedback with static memoryless nonlinearities, and regulating the $\ell_\infty$ gain through the design of a state-feedback controller. Finite incremental gains provide a quantitative measure of robustness for trajectories,

Goro Shibata, Keisuke Ikeda, Takeshi Seki, Shoya Sakamoto

Among magnetic thin films with perpendicular magnetic anisotropy (PMA), $L1_0$-ordered FePt has attracted significant attention because of its exceptionally strong PMA. However, the microscopic origin of its strong PMA has not been elucidated experimentally. We have investigated the contribution of the Fe $3d$ electrons to its magnetic anisotropy energy by a

Zeyi Chen, Ariel Neufeld, Qikun Xiang

We develop an estimator-based stochastic fixed-point framework for approximately computing the 2-Wasserstein barycenter of continuous, non-parametric probability measures. Notably, we provide the first rigorous convergence analysis for implementable estimator-based stochastic extensions of the fixed-point iterative scheme proposed by \'Alvarez-Esteban, del B

Simone Di Gregorio, Francesco Iafrate

We take into consideration generalization bounds for the problem of the estimation of the drift component for ergodic stochastic differential equations, when the estimator is a ReLU neural network and the estimation is non-parametric with respect to the statistical model. We show a practical way to enforce the theoretical estimation procedure, enabling infer

Wen Fan, Haoran Li, Dandan Zhang

Contact-rich manipulation in unstructured environments demands precise, multimodal perception to enable robust and adaptive control. Vision-based tactile sensors (VBTSs) have emerged as an effective solution; however, conventional VBTSs often face challenges in achieving compact, multi-modal functionality due to hardware constraints and algorithmic complexit

Guo Chen, Bo Ning, Jianhua Tu

The independence polynomial of a graph is termed {\it stable} if all its roots are located in the left half-plane $\{z \in \mathbb{C} : \mathrm{Re}(z) \leq 0\}$, and the graph itself is also referred to as stable. Brown and Cameron (Electron. J. Combin. 25(1) (2018) \#P1.46) proved that the complete bipartite graph $K_{1,n}$ is stable and posed the question:

Zheng Wang

Fine-grained bird image classification (FBIC) is not only of great significance for ecological monitoring and species identification, but also holds broad research value in the fields of image recognition and fine-grained visual modeling. Compared with general image classification tasks, FBIC poses more formidable challenges: 1) the differences in species si

Xiaoyu Wu, Yifei Pang, Terrance Liu, Zhiwei Steven Wu

Large Language Models are typically trained on datasets collected from the web, which may inadvertently contain harmful or sensitive personal information. To address growing privacy concerns, unlearning methods have been proposed to remove the influence of specific data from trained models. Of these, exact unlearning -- which retrains the model from scratch

Yilun Kong, Guozheng Ma, Qi Zhao, Haoyu Wang

Despite recent advancements in offline multi-task reinforcement learning (MTRL) have harnessed the powerful capabilities of the Transformer architecture, most approaches focus on a limited number of tasks, with scaling to extremely massive tasks remaining a formidable challenge. In this paper, we first revisit the key impact of task numbers on current MTRL m

Yu-Hsuan Lin, Qian-Hui Chen, Yi-Jie Cheng, Jia-Ren Zhang

Recent advancements in large language models (LLMs) have enhanced natural-language reasoning. However, their limited parametric memory and susceptibility to hallucination present persistent challenges for tasks requiring accurate, context-based inference. To overcome these limitations, an increasing number of studies have proposed leveraging external knowled

Wei-Han Tan, Wen-Ying Liu, Hong-Zhou Xi, Hua-Xing Chen

We apply the QCD sum rule method to systematically study excited light meson operators and calculate their decay constants. These operators are constructed by explicitly adding one covariant derivative to the quark-antiquark pair. In total, twelve such operators are constructed, among which ten are subjected to detailed numerical analyses. The considered qua

Maciej Wielgosz, Simon Berg, Heikki Korpunen, Stephan Hoffmann

This paper presents a deep learning-based framework for classifying forestry operations from dashcam video footage. Focusing on four key work elements - crane-out, cutting-and-to-processing, driving, and processing - the approach employs a 3D ResNet-50 architecture implemented with PyTorchVideo. Trained on a manually annotated dataset of field recordings, th

J. Blaha, L. Xu, M. La Mantia

We study experimentally the starting vortices shed by airfoils accelerating uniformly from rest in superfluid helium-4 (He II). The vortices behave apparently as if they were moving in a classical Newtonian fluid, such as air or water. Specifically, the starting vortex positions obtained from the experimental data are found to be very close to those computed

J. M. Palencia, Paloma Morilla, Sung Kei Li, J. M. Diego

We investigate the strong gravitational lensing properties of fuzzy dark matter (FDM) halos, focusing on the magnification properties near radial critical curves (CCs). Using simulated lenses we compute magnification maps for a range of axion masses and halo configurations. We show that FDM produces enhanced central magnification and secondary CCs that are n

Xinglin Wang, Yiwei Li, Shaoxiong Feng, Peiwen Yuan

Test-Time Scaling (TTS) improves the performance of Large Language Models (LLMs) by using additional inference-time computation to explore multiple reasoning paths through search. Yet how to allocate a fixed rollout budget most effectively during search remains underexplored, often resulting in inefficient use of compute at test time. To bridge this gap, we

Yichi Zhang, Gongwei Chen, Jun Zhu, Jia Wan

Visual grounding requires large and diverse region-text pairs. However, manual annotation is costly and fixed vocabularies restrict scalability and generalization. Existing pseudo-labeling pipelines often overfit to biased distributions and generate noisy or redundant samples. Through our systematic analysis of data quality and distributional coverage, we fi

Md Intisar Chowdhury, Kittinun Aukkapinyo, Hiroshi Fujimura, Joo Ann Woo

In this paper, we propose a Grid-based Local and Global Area Transcription (Grid-LoGAT) system for Video Question Answering (VideoQA). The system operates in two phases. First, extracting text transcripts from video frames using a Vision-Language Model (VLM). Next, processing questions using these transcripts to generate answers through a Large Language Mode

Fabio Briscese, Francesco Calogero, Farrin Payandeh

In this paper we discuss some remarkable properties of the autonomous system of 2 first-order Ordinary Differential Equations (ODEs), which equates the derivatives $\dot{x}_n(t)$ ($n = 1, 2$) of the 2 dependent variables $x_n(t)$ to the ratios of polynomials (with constant coefficients) in the 2 variables $x_n (t)$: each of the 2 (a priori different) polynom

Yuanfu Wang, Pengyu Wang, Chenyang Xi, Bo Tang

Modern language models often rely on Reinforcement Learning from Human Feedback (RLHF) to encourage safe behaviors. However, they remain vulnerable to adversarial attacks due to three key limitations: (1) the inefficiency and high cost of human annotation, (2) the vast diversity of potential adversarial attacks, and (3) the risk of feedback bias and reward h

Fabio Punzo

We study uniqueness for solutions to the Cauchy problem associated with the parabolic Schr\"odinger equation on complete noncompact Riemannian manifolds, under suitable integral conditions on the solution. We show that, under suitable assumptions on the potential V, the required integrability condition can be significantly relaxed compared to the case withou

Zhao-Qi-Zhi Han, Xiao-Hua Wang, Jin-Peng Li, Bo-Wen Liu

Long-wavelength infrared band exhibits significant utility in thermal signature acquisition and molecular spectral analysis, among other applications. The up-conversion detection technique enables effective signal transduction into the detection bandwidth of silicon-based photodetectors, thereby facilitating high-sensitivity photonic measurements. We realize

N. A. Moroz, K. S. Tikhonov, L. V. Gerasimov, A. D. Manukhova

The permutation symmetry is a fundamental attribute of the collective wavefunction of indistinguishable particles. It makes a difference for the behavior of collective systems having different quantum statistics but existing in the same environment. Here we show that for some specific quantum conjugation between the spin and spatial degrees of freedom the in

Vardhan Shorewala, Shivam Shorewala

This paper introduces a unified approach to cluster refinement and anomaly detection in datasets. We propose a novel algorithm that iteratively reduces the intra-cluster variance of N clusters until a global minimum is reached, yielding tighter clusters than the standard k-means algorithm. We evaluate the method using intrinsic measures for unsupervised lear

Zexin Fu, Riccardo Tedeschi, Gianmarco Ottavi, Nils Wistoff

Open-source RISC-V cores are increasingly demanded in domains like automotive and space, where achieving high instructions per cycle (IPC) through superscalar and out-of-order (OoO) execution is crucial. However, high-performance open-source RISC-V cores face adoption challenges: some (e.g. BOOM, Xiangshan) are developed in Chisel with limited support from i

Anum Afzal, Florian Matthes, Gal Chechik, Yftah Ziser

We investigate whether the success of a zero-shot Chain-of-Thought (CoT) process can be predicted before completion. We discover that a probing classifier, based on LLM representations, performs well \emph{even before a single token is generated}, suggesting that crucial information about the reasoning process is already present in the initial steps represen

Jan Kuske, Almudena Arcones, Moritz Reichert

Heavy elements are synthesized by the r-process in neutron star mergers and potentially in rare supernovae linked to strong magnetic fields. Expensive hydrodynamic simulations of these extreme environments are usually post-processed to calculate the nucleosynthesis. In contrast, here we follow a site-independent approach based on three key parameters: electr

Roger Ferrod, Cássio F. Dantas, Luigi Di Caro, Dino Ienco

Multi-modal RGB and Depth (RGBD) data are predominant in many domains such as robotics, autonomous driving and remote sensing. The combination of these multi-modal data enhances environmental perception by providing 3D spatial context, which is absent in standard RGB images. Although RGBD multi-modal data can be available to train computer vision models, acc

Stepan Shabalin, Ayush Panda, Dmitrii Kharlapenko, Abdur Raheem Ali

Sparse autoencoders are a promising new approach for decomposing language model activations for interpretation and control. They have been applied successfully to vision transformer image encoders and to small-scale diffusion models. Inference-Time Decomposition of Activations (ITDA) is a recently proposed variant of dictionary learning that takes the dictio

Prakash Pandey, Sudhir K. Pandey

Several topological electronic materials have been theoretically predicted, leading to a comprehensive catalog systematically characterized by their band crossings. Researchers have attempted to experimentally verify the topological nature of some materials from the present catalogs, but not all efforts have yielded positive results. Here, we introduce a pos

Abhinav Bitragunta, Hareshkumar Jadav, Ranveer Singh

We introduce a method for constructing larger families of connected cospectral graphs from two given cospectral families of sizes $p$ and $q$. The resulting family size depends on the Cartesian primality of the input graphs and can be one of $pq$, $p + q - 1$, or $\max(p, q)$, based on the strictness of the applied conditions. Under the strictest condition,

Xianglong Yan, Zhiteng Li, Tianao Zhang, Haotong Qin

Large language models (LLMs) have demonstrated remarkable performance, but their long-context reasoning remains constrained by the excessive memory required for the Key-Value (KV) cache. This makes KV cache compression a critical step toward efficient long-context inference. Recent methods have explored low-rank techniques to reduce the hidden size of the KV

Jinyang Li, Jianyu Wang, Wenchi Cheng, Yudong Fang

In this paper, we enhance the omnidirectional coverage performance of tri-directional coil-based magnetic induction communication (TC-MIC) and reduce the pathloss with a joint transmit and receive magnetic beamforming method. An iterative optimization algorithm incorporating the transmit current vector and receive weight matrix is developed to minimize the p

Sihan Tan, Taro Miyazaki, Kazuhiro Nakadai

Sign Language Translation (SLT) aims to convert sign language (SL) videos into spoken language text, thereby bridging the communication gap between the sign and the spoken community. While most existing works focus on translating a single sign language into a single spoken language (one-to-one SLT), leveraging multilingual resources could mitigate low-resour

Qianqian Zhang, Jiajia Liao, Heting Ying, Yibo Ma

Language agents powered by large language models (LLMs) have demonstrated remarkable capabilities in understanding, reasoning, and executing complex tasks. However, developing robust agents presents significant challenges: substantial engineering overhead, lack of standardized components, and insufficient evaluation frameworks for fair comparison. We introdu