March 2025 arXiv papers — page 166

Audun Myers, Max Vargas, Sinan G. Aksoy, Cliff Joslyn

In this work we study various Retrieval Augmented Regeneration (RAG) approaches to gain an understanding of the strengths and weaknesses of each approach in a question-answering analysis. To gain this understanding we use a case-study subset of the Global Database of Events, Language, and Tone (GDELT) dataset as well as a corpus of raw text scraped from the

Batukhan Azheev, Nikita Tselousov

We develop methods for systematic construction of superintegrable polynomials in matrix/eigenvalue models. Our consideration is based on a tight connection of superintegrable property of Gaussian Hermitian model and $W_{1 + \infty}$ algebra in Fock representation. Motivated by this example, we propose a set of assumptions that may allow one to recover superi

Holger Lange, Michael Hauhs

Small, forested catchments are prototypes of terrestrial ecosystems and have been studied in several disciplines of environmental sciences since several decades. Time series of water and matter fluxes and nutrient concentrations from these systems exhibit a bewildering diversity of spatio-temporal patterns, indicating the intricate nature of processes acting

Denver-James Logan Marchment

Let ${p > 2}$ be an odd prime and ${G = SL_2(\mathbb{F}_p)}$. Denote the subgroup of upper triangular matrices as $B$. Finally, let ${\mathbb{F}}$ be an algebraically closed field of characteristic ${p}$. The Green correspondence gives a bijection between the non-projective indecomposable ${\mathbb{F}[G]}$ modules and non-projective indecomposable ${\mathbb{

Zijun Liao, Jinbiao Chen, Debing Wang, Zizhen Zhang

Neural Combinatorial Optimization (NCO) has emerged as a promising approach for NP-hard problems. However, prevailing RL-based methods suffer from low sample efficiency due to sparse rewards and underused solutions. We propose Best-anchored and Objective-guided Preference Optimization (BOPO), a training paradigm that leverages solution preferences via object

Martin Kjøllesdal Johnsrud, Ramin Golestanian

Interactions between active particles may be non-reciprocal, breaking action-reaction symmetry and leading to novel physics not observed in equilibrium systems. The non-reciprocalCahn-Hilliard (NRCH) model is a phenomenological model that captures the large-scale effects of non-reciprocity in conserved, phase-separating systems. In this work, we explore the

Tianyu Chen, Yasi Zhang, Zhendong Wang, Ying Nian Wu

Diffusion models have achieved remarkable success in generating high-resolution, realistic images across diverse natural distributions. However, their performance heavily relies on high-quality training data, making it challenging to learn meaningful distributions from corrupted samples. This limitation restricts their applicability in scientific domains whe

Deijany Rodriguez Linares, Oksana Moryakova, Håkan Johansson

This paper introduces a sampling frequency offset (SFO) estimation method based on the Farrow structure, which is typically utilized for the SFO compensation and thereby enables a reduction of the implementation complexity of the SFO estimation. The proposed method is implemented in the time domain and works for arbitrary bandlimited signals, thus with no ad

Raphael Gerlach, Sören von der Gracht

In this article, we investigate symmetry properties of distributed systems of mobile robots. We consider a swarm of $n\in\mathbb{N}$ robots in the $\mathcal{OBLOT}$ model and analyze their collective $\mathcal{F}$sync dynamics using of equivariant dynamical systems theory. To this end, we show that the corresponding evolution function commutes with rotationa

Konstantinos Vergopoulos, Mark Niklas Müller, Martin Vechev

Code Agent development is an extremely active research area, where a reliable performance metric is critical for tracking progress and guiding new developments. This demand is underscored by the meteoric rise in popularity of SWE-Bench. This benchmark challenges code agents to generate patches addressing GitHub issues given the full repository as context. Th

Jen-tse Huang, Jiantong Qin, Jianping Zhang, Youliang Yuan

This research investigates both explicit and implicit social biases exhibited by Vision-Language Models (VLMs). The key distinction between these bias types lies in the level of awareness: explicit bias refers to conscious, intentional biases, while implicit bias operates subconsciously. To analyze explicit bias, we directly pose questions to VLMs related to

Fabrizio Barone, Marco Casazza

This study proposes a new approach to the interpretation of Greek Doric-style temples, based on the integration of its tangible and intangible dimensions as a cultural heritage asset. Rooted on the Greek concept of techne, the work considers a unifying design principle, integrating both structural and functional aspects within the architectural style. A mult

Vojtěch Kala, Jiří Fadrný, Michal Neset, Jan Bílek

Randomness is a key feature of quantum physics. Heisenberg's uncertainty principle reveals the existence of an intrinsic noise, usually explored through Gaussian squeezed states. Due to their insufficiency for quantum advantage, the focus is currently shifting towards genuinely quantum non-Gaussian states. However, while genuine quantum behavior comes natura

Aidan Backus

We prove an inversion formula for the exterior $k$-plane transform. As a consequence, we show that if $m < k$ then an $m$-current in $\mathbf R^n$ can be reconstructed from its projections onto $\mathbf R^k$, which proves a conjecture of Solomon.

Yuxiao Qu, Matthew Y. R. Yang, Amrith Setlur, Lewis Tunstall

Training models to effectively use test-time compute is crucial for improving the reasoning performance of LLMs. Current methods mostly do so via fine-tuning on search traces or running RL with 0/1 outcome reward, but do these approaches efficiently utilize test-time compute? Would these approaches continue to scale as the budget improves? In this paper, we

Vilas Winstein

Folklore belief holds that metastable wells in low-temperature statistical mechanics models exhibit high-temperature behavior. We make this rigorous in the exponential random graph model (ERGM) through the lens of concentration of measure. We make use of the supercritical (low-temperature) metastable mixing which was recently proven by Bresler, Nagaraj, and

Khalid Baadi

In this paper, we consider second order degenerate parabolic equations with complex, measurable, and time-dependent coefficients. The degenerate ellipticity is dictated by a spatial $A_2$-weight. We prove that having a generalized fundamental solution with upper Gaussian bounds is equivalent to Moser's $L^2$-$L^\infty$ estimates for local weak solutions. In

Habibur Rahaman, Atri Chatterjee, Swarup Bhunia

Rapid adoption of AI technologies raises several major security concerns, including the risks of adversarial perturbations, which threaten the confidentiality and integrity of AI applications. Protecting AI hardware from misuse and diverse security threats is a challenging task. To address this challenge, we propose SAMURAI, a novel framework for safeguardin

Nithin Raveendran, David Declercq, Bane Vasić

Quantum error correction (QEC) is critical for practical realization of fault-tolerant quantum computing, and recently proposed families of quantum low-density parity-check (QLDPC) code are prime candidates for advanced QEC hardware architectures and implementations. This paper focuses on the finite-length QLDPC code design criteria, specifically aimed at co

Aaron Zoll, Benjamin Grimmer

This paper proposes a universal algorithm for convex minimization problems of the composite form $g_0(x)+h(g_1(x),\dots, g_m(x)) + u(x)$. We allow each $g_j$ to independently range from being nonsmooth Lipschitz to smooth, from convex to strongly convex, described by notions of H\"older continuous gradients and uniform convexity. Note that, although the obje

Linqi Zhou, Stefano Ermon, Jiaming Song

Diffusion models and Flow Matching generate high-quality samples but are slow at inference, and distilling them into few-step models often leads to instability and extensive tuning. To resolve these trade-offs, we propose Inductive Moment Matching (IMM), a new class of generative models for one- or few-step sampling with a single-stage training procedure. Un

Yingdong Guan, Abhinava Chatterjee, Trace Bivens, Seng Huat Lee

Topological semimetals, particularly Weyl semimetals (WSMs), are crucial platforms for exploring emergent quantum phenomena due to their unique electronic structures and potential to transition into various topological phases. In this study, we report the discovery of a ferromagnetic (FM) type-II WSM in Mn(Bi1-xSbx)4Te7, which exhibits a remarkable three-dim

Canyi Chen, Nan Qiao, Liping Zhu

This paper addresses the problem of efficiently classifying high-dimensional data over decentralized networks. Penalized support vector machines (SVMs) are widely used for high-dimensional classification tasks. However, the double nonsmoothness of the objective function poses significant challenges in developing efficient decentralized learning methods. Many

J. Xu, J. Kim, B. Lenardo, C. E. Dahl

The response of liquid xenon to various types of ionizing radiation has been extensively studied theoretically and experimentally. Recent progress in direct detection dark matter experiments highlights the significance of composite events, where multiple particles interact with xenon simultaneously and generate overlapping ionization signatures. In these eve

Thibaut Loiseau, Guillaume Bourmaud, Vincent Lepetit

Pre-training techniques have greatly advanced computer vision, with CroCo's cross-view completion approach yielding impressive results in tasks like 3D reconstruction and pose regression. However, cross-view completion is ill-posed in non-covisible regions, limiting its effectiveness. We introduce Alligat0R, a novel pre-training approach that replaces cross-

Jennifer Coulter, Bogdan Rajkov, Michele Simoncelli

Non-diffusive, fluid-like transport of charge and heat has been observed in several materials, raising the question of whether they can emerge simultaneously and how they are related to bi-component electron-phonon fluids. Here we introduce a first-principles theory and computational framework to quantitatively describe these phenomena from atomistic to cont

Eldar Knar

Current research funding systems are subject to structural imbalances, where formal criteria such as the H-index and the number of publications dominate over the potential and actual scientific prospects of researchers. This leads to the suppression of potential breakthrough research directions and limited access to grants for stochastic (innovative) researc

C. W. Sandbo Chang, Arjan F. Van Loo, Chih-Chiao Hung, Yu Zhou

We present a Josephson traveling-wave parametric amplifier (JTWPA) based on a low-loss coplanar lumped-element waveguide architecture. By employing open-stub capacitors and Manhattan-pattern junctions, our device achieves an insertion loss below 1~dB up to 12~GHz. We introduce windowed sinusoidal modulation for phase matching, demonstrating that a smooth tra

Jason Milionis, Jens Ernstberger, Joseph Bonneau, Scott Duke Kominers

We introduce the first formal model capturing the elicitation of unverifiable information from a party (the "source") with implicit signals derived by other players (the "observers"). Our model is motivated in part by applications in decentralized physical infrastructure networks (a.k.a. "DePIN"), an emerging application domain in which physical services (e.

Hossein Karami, Antony Thomas, Fulvio Mastrogiovanni

In this paper, we present an approach for integrated task and motion planning based on an AND/OR graph network, which is used to represent task-level states and actions, and we leverage it to implement different classes of task and motion planning problems (TAMP). Several problems that fall under task and motion planning do not have a predetermined number of

Yangzhe Kong, Daeun Song, Jing Liang, Dinesh Manocha

We present a novel method, AutoSpatial, an efficient approach with structured spatial grounding to enhance VLMs' spatial reasoning. By combining minimal manual supervision with large-scale Visual Question-Answering (VQA) pairs auto-labeling, our approach tackles the challenge of VLMs' limited spatial understanding in social navigation tasks. By applying a hi

Samuel Ferino, Rashina Hoda, John Grundy, Christoph Treude

Following the rise of large language models (LLMs), many studies have emerged in recent years focusing on exploring the adoption of LLM-based tools for software development by novice developers: computer science/software engineering students and early-career industry developers with two years or less of professional experience. These studies have sought to u

Fateme Jamshidi, Mohammad Shahverdikondori, Negar Kiyavash

We study multi-armed bandits under network interference, where each unit's reward depends on its own treatment and those of its neighbors in a given graph. This induces an exponentially large action space, making standard approaches computationally impractical. We propose a novel algorithm that uses the local graph structure to minimize regret. We derive a g

Céline Hocquette, Andrew Cropper

Recent inductive logic programming (ILP) approaches learn optimal hypotheses. An optimal hypothesis minimises a given cost function on the training data. There are many cost functions, such as minimising training error, textual complexity, or the description length of hypotheses. However, selecting an appropriate cost function remains a key question. To addr

Maximilian Seidler, Alexander Krause, Peter Ulbrich

Containerization has become a ubiquitous tool in software development. Due to its numerous benefits, including platform interoperability and secure execution of untrusted third-party code, this technology is a boon to industrial automation, promising to provide aid for their inherent challenges - except one, which is interaction with physical devices. Unfort

Duy Phuong Nguyen, J. Pablo Munoz, Tanya Roosta, Ali Jannesari

Federated Learning (FL) enables collaborative learning across distributed clients while preserving data privacy. However, FL faces significant challenges when dealing with heterogeneous data distributions, which can lead to suboptimal global models that fail to generalize across diverse clients. In this work, we propose a novel framework designed to tackle t

Huiyang Shao, Xin Xia, Yuhong Yang, Yuxi Ren

Diffusion models have achieved remarkable success across various domains. However, their slow generation speed remains a critical challenge. Existing acceleration methods, while aiming to reduce steps, often compromise sample quality, controllability, or introduce training complexities. Therefore, we propose RayFlow, a novel diffusion framework that addresse

Paul Boniol, Donato Tiano, Angela Bonifati, Themis Palpanas

With the exponential growth of time series data across diverse domains, there is a pressing need for effective analysis tools. Time series clustering is important for identifying patterns in these datasets. However, prevailing methods often encounter obstacles in maintaining data relationships and ensuring interpretability. We present Graphint, an innovative

Pritam Ganguly, Jayanta Sarkar

This article presents the $L^p$-Heisenberg--Pauli--Weyl uncertainty inequality for the group Fourier transform on a class of two-step nilpotent Lie groups, specifically the M\'etivier groups. This inequality quantitatively demonstrates that on M\'etivier groups, a nonzero function and its group Fourier transform cannot both be sharply localized. The proof pr

Yuqi Ding, Elisa Donati, Haobo Li, Hadi Heidari

Wearable health devices have a strong demand in real-time biomedical signal processing. However traditional methods often require data transmission to centralized processing unit with substantial computational resources after collecting it from edge devices. Neuromorphic computing is an emerging field that seeks to design specialized hardware for computing s

Haoran Li, Junfeng Hu

Large Language Models (LLMs) have demonstrated remarkable capabilities in various tasks, yet still produce errors in domain-specific tasks. To further improve their performance, we propose KSOD (Knowledge Supplement for LLMs On Demand), a novel framework that empowers LLMs to improve their capabilities with knowledge-based supervised fine-tuning (SFT). KSOD

Lisa Johanna Schumacher, Mauricio Bustamante, Matteo Agostini, Foteini Oikonomou

Decades of progress have culminated in first light for high-energy neutrino astronomy: the identification of the first astrophysical sources of TeV-PeV neutrinos by the IceCube neutrino telescope, the active galactic nuclei NGC 1068 and TXS 0506+056. Today, the prospect of going beyond first light to build high-energy neutrino astronomy in earnest by discove

Nina Moorman, Michelle Zhao, Matthew B. Luebbers, Sanne Van Waveren

In human-robot interactions, human and robot agents maintain internal mental models of their environment, their shared task, and each other. The accuracy of these representations depends on each agent's ability to perform theory of mind, i.e. to understand the knowledge, preferences, and intentions of their teammate. When mental models diverge to the extent

Michael-Andrei Panaitescu-Liess, Pankayaraj Pathmanathan, Yigitcan Kaya, Zora Che

As the capabilities of large language models (LLMs) continue to expand, their usage has become increasingly prevalent. However, as reflected in numerous ongoing lawsuits regarding LLM-generated content, addressing copyright infringement remains a significant challenge. In this paper, we introduce PoisonedParrot: the first stealthy data poisoning attack that

Jeffrey Stopple

Study of the level curve for the real part of $\eta(s)=0$ with $\eta(s)=\pi^{-s/2}\Gamma(s/2)\zeta^\prime(s)$ gives a new classification of the zeros of $\zeta(s)$ and of $\zeta^\prime(s)$. We conjecture that for type 2 zeros, $\liminf (\beta^\prime -1/2)\log\gamma^\prime = 0$ if and only if $\liminf (\gamma^+-\gamma^-)\log \gamma^\prime=0$, and reduce the c

Marco Bresciani, Manuel Friedrich

We study a variational model in nonlinear elasticity allowing for cavitation which penalizes both the volume and the perimeter of the cavities. Specifically, we investigate the approximation (in the sense of {\Gamma}-convergence) of the energy by means of functionals defined on perforated domains. Perforations are introduced at flaw points where singularitie

Michael Mitzenmacher, Rana Shahout

Queueing systems present many opportunities for applying machine-learning predictions, such as estimated service times, to improve system performance. This integration raises numerous open questions about how predictions can be effectively leveraged to improve scheduling decisions. Recent studies explore queues with predicted service times, typically aiming

Laur Järv, Dmitri Kraiko

In the phase space perspective, scalar field slow roll inflation is described by a heteroclinic orbit from a saddle type fixed point to a final attractive point. In many models the saddle point resides in the scalar field asymptotics, and thus for a comprehensive view of the dynamics a global phase portrait is necessary. For this task, in the literature one

Ko Honda, Roman Krutowski, Yin Tian, Tianyu Yuan

Given a smooth closed $n$-manifold $M$ and a $\kappa$-tuple of basepoints $\boldsymbol{q}\subset M$, we define a Morse-type $A_\infty$-algebra $CM_{-*}(\Omega(M,\boldsymbol{q}))$, called the based multiloop $A_\infty$-algebra, as a graded generalization of the braid skein algebra due to Morton and Samuelson. For example, when $M=T^2$ the braid skein algebra

Bhargab Das, Aprameyo Pal

In this article, we investigate factorization problems for twisted triple product Galois representations over real quadratic fields, arising from families of Hilbert cusp forms. Specifically, we address the factorization in two distinct settings determined by the order of vanishing of associated $L$-unctions at their central critical values-namely, the rank

Zhao-Heng Yin, Changhao Wang, Luis Pineda, Krishna Bodduluri

We introduce Geometric Retargeting (GeoRT), an ultrafast, and principled neural hand retargeting algorithm for teleoperation, developed as part of our recent Dexterity Gen (DexGen) system. GeoRT converts human finger keypoints to robot hand keypoints at 1KHz, achieving state-of-the-art speed and accuracy with significantly fewer hyperparameters. This high-sp

Maurice Ongala, Ruth Kiraka, Jyoti Choundrie, Javan Okello

This protocol outlines a scoping review designed to systematically map the existing body of evidence on AI-enabled knowledge sharing in resource-limited non-profit healthcare organizations. The review aims to investigate how such technologies enhance collaboration and decision-making, particularly in the context of reduced external support following the cess

Loni Hagen, Ly Dinh, Golfo Alexopoulos, Lingyao Li

On February 7, 2024, Russian President Vladimir Putin gave a two-hour interview with conservative political commentator, Tucker Carlson. This study investigated the impact of the Carlson- Putin interview on the US X audience. We proposed a framework of social media impact using machine learning (ML) and natural language processing (NLP) by measuring changes

Zhenyu Li, Kehai Chen, Yunfei Long, Xuefeng Bai

Large Language Models (LLMs) have demonstrated remarkable instruction-following capabilities across various applications. However, their performance in multilingual settings lacks systematic investigation, with existing evaluations lacking fine-grained constraint analysis across diverse linguistic contexts. We introduce XIFBench, a comprehensive constraint-b

Tucker Allen, Barry Y. Li, Tim Duong, Kajsa Williams

An efficient mixed deterministic/sparse-stochastic plane-wave approach is developed for bandstructure calculations of large supercell periodic generalized-Kohn-Sham density functional theory, for any hybrid-exchange density functional. The method works for very large elementary cells and supercells, and we benchmark it on covalently bonded solids and molecul

Oskar A. Sultanov

The effect of multiplicative white noise on the resonance capture in non-isochronous systems with time-decaying pumping is investigated. It is assumed that the intensity of perturbations decays with time, and its frequency is asymptotically constant. The occurrence of attractive solutions with an amplitude close to the resonant value and a phase synchronized

Yingzhe Peng, Gongrui Zhang, Miaosen Zhang, Zhiyuan You

Enhancing reasoning in Large Multimodal Models (LMMs) faces unique challenges from the complex interplay between visual perception and logical reasoning, particularly in compact 3B-parameter architectures where architectural constraints limit reasoning capacity and modality alignment. While rule-based reinforcement learning (RL) excels in text-only domains,

Clément Chadebec, Onur Tasar, Sanjeev Sreetharan, Benjamin Aubin

In this paper, we introduce Latent Bridge Matching (LBM), a new, versatile and scalable method that relies on Bridge Matching in a latent space to achieve fast image-to-image translation. We show that the method can reach state-of-the-art results for various image-to-image tasks using only a single inference step. In addition to its efficiency, we also demon

Frederik J. Thomsen, Johan L. A. Dubbeldam

Adaptive therapy is a recent paradigm in cancer treatment aiming at indefinite, safe containment of the disease when cure is judged unattainable. In modeling this approach, inherent limitations arise due to the structure of the vector fields and the bounds imposed by toxic side-effects of the drug. In this work we analyze these limitations in a minimal class

Lyla Choi, Adam Burrows, David Vartanyan

In this paper, we examine the neutrino signals from 24 initially non-rotating, three-dimensional core-collapse supernova (CCSN) simulations carried to late times. We find that not only does the neutrino luminosity signal encode information about each stage of the CCSN process, but that the monotonic dependence of the luminosity peak height with compactness e

Igor Kortchemski, Leonard Vetter

The goal of this note is to study the geometry of large size-conditioned Bienaym\'e trees whose offspring distribution is subcritical, belongs to the domain of attraction of a stable law of index $\alpha=1$ and satisfies a local regularity assumption. We show that a condensation phenomenon occurs: one unique vertex of macroscopic degree emerges, and its heig

M. Stalport, A. Mortier, M. Cretignier, J. A. Egger

We report on the validation and characterisation of two transiting planets around TOI-1453, a K-dwarf star in the TESS northern continuous viewing zone. In addition to the TESS data, we used ground-based photometric, spectroscopic, and high-resolution imaging follow-up observations to validate the two planets. We obtained 100 HARPS-N high-resolution spectra

Anastasiia Grishina, Vadim Liventsev, Aki Härmä, Leon Moonen

Program synthesis with Large Language Models (LLMs) suffers from a "near-miss syndrome": the generated code closely resembles a correct solution but fails unit tests due to minor errors. We address this with a multi-agent framework called Synthesize, Execute, Instruct, Debug, and Repair (SEIDR). Effectively applying SEIDR to instruction-tuned LLMs requires d

Qasim Khadim, Peter Manzl, Emil Kurvinen, Aki Mikkola

The precision, stability, and performance of lightweight high-strength steel structures in heavy machinery is affected by their highly nonlinear dynamics. This, in turn, makes control more difficult, simulation more computationally intensive, and achieving real-time autonomy, using standard approaches, impossible. Machine learning through data-driven, physic

Kaida Wu, Peihao Xiang, Chaohao Lin, Ou Bai

To enhance lifting-load estimation accuracy in industrial upper-limb assistive exoskeletons, this study proposes a machine learning-based approach using insole pressure sensors. Unlike traditional methods that rely on electromyography (EMG), force sensors, or posture data, insole pressure sensors provide a non-invasive, posture-independent, and stable soluti

H. L. Devereux, C. Cockrell, A. M. Elena, Ian Bush

Modelling has become a third distinct line of scientific enquiry, alongside experiments and theory. Molecular dynamics (MD) simulations serve to interpret, predict and guide experiments and to test and develop theories. A major limiting factor of MD simulations is system size and in particular the difficulty in handling, storing and processing trajectories o

Hisham A. Alyahya, Haidar Khan, Yazeed Alnumay, M Saiful Bari

We introduce ZeroSumEval, a dynamic, competition-based, and evolving evaluation framework for Large Language Models (LLMs) that leverages competitive games. ZeroSumEval encompasses a diverse suite of games, including security challenges (Capture the Flag), classic board games (chess), and knowledge tests (MathQuiz). These games are designed to evaluate a ran

Shingo Takeuchi

In this study, we first heuristically construct the charges corresponding to the chiral transformation associated with the large U(1) gauge symmetry. We refer to these as the large chiral charges, and to the chiral transformation they generate as the large chiral transformations. Then, showing that these large chiral charges can be obtained based on Noether'

Yuzhe Qin, Cheng Wang

In this paper, we propose and analyze a second order accurate (in both time and space) numerical scheme for the Poisson-Nernst-Planck-Navier-Stokes system, which describes the ion electro-diffusion in fluids. In particular, the Poisson-Nernst-Planck equation is reformulated as a non-constant mobility gradient flow in the Energetic Variational Approach. The m

Aixia Zhang, Qingyun Zhang, Zhiyi Chen, Yong Wu

Screened spherical wave (SSW) of the Hankel function features the complete, minimal and short-ranged basis set, presenting a compact representation for electronic systems. In this work, we report the implementation of full-potential (FP) SSW based tight-binding linearized Muffin-Tin orbital (TB-LMTO) for all-electron density functional theory (DFT), and prov

Thibaud Leteno, Michael Perrot, Charlotte Laclau, Antoine Gourru

Group fairness is a central research topic in text classification, where reaching fair treatment between sensitive groups (e.g., women and men) remains an open challenge. We propose an approach that extends the use of the Wasserstein Dependency Measure for learning unbiased neural text classifiers. Given the challenge of distinguishing fair from unfair infor

Zhangquan Chen, Xufang Luo, Dongsheng Li

Visual understanding is inherently intention-driven - humans selectively focus on different regions of a scene based on their goals. Recent advances in large multimodal models (LMMs) enable flexible expression of such intentions through natural language, allowing queries to guide visual reasoning processes. Frameworks like Visual Chain-of-Thought have demons

Purvi Agrawal, Vikas Joshi, Bharati Patidar, Ankur Gupta

In this paper, we present a novel approach to developing an English Automatic Speech Recognition (ASR) system that can effectively handle Hindi queries, without compromising its performance on English. We propose a novel acoustic model (AM), referred to as SplitHead with Attention (SHA) model, features shared hidden layers across languages and language-speci

Irina Ya. Aref'eva, Ali Hajilou, Alexander Nikolaev, Pavel Slepov

We investigate the influence of a magnetic field on the running coupling constant for a heavy-quark model in a bottom-up holographic approach. To achieve this, we employ a magnetized Einstein-Maxwell-dilaton background that captures the essential features of heavy quark dynamics. Similar to the light-quark model, the running coupling $\alpha$ for heavy quark

Zhongwei Chen, Zhao-Xu Yang, Hai-Jun Rong, Jiawei Lang

Traditional supervised drone-view geo-localization (DVGL) methods heavily depend on paired training data and encounter difficulties in learning cross-view correlations from unpaired data. Moreover, when deployed in a new domain, these methods require obtaining the new paired data and subsequent retraining for model adaptation, which significantly increases c

Justus-Jonas Erker, Nils Reimers, Iryna Gurevych

Decomposition-based multi-hop retrieval methods rely on many autoregressive steps to break down complex queries, which breaks end-to-end differentiability and is computationally expensive. Decomposition-free methods tackle this, but current decomposition-free approaches struggle with longer multi-hop problems and generalization to out-of-distribution data. T

Takeru Inoue, Ryusuke Miyamoto

Instance shadow detection is the task of detecting pairs of shadows and objects, where existing methods first detect shadows and objects independently, then associate them. This paper introduces FastInstShadow, a method that enhances detection accuracy through a query-based architecture featuring an association transformer decoder with two dual-path transfor

Weize Li, Yunhao Du, Qixiang Yin, Zhicheng Zhao

Referring Multi-Object Tracking (RMOT) aims to track multiple objects specified by natural language expressions in videos. With the recent significant progress of one-stage methods, the two-stage Referring-by-Tracking (RBT) paradigm has gradually lost its popularity. However, its lower training cost and flexible incremental deployment remain irreplaceable. R

Xiuyuan Li, Matteo Pagliero, Gábor Szabó

This article concerns the structure of $\mathrm{C}^{\ast}$-algebraic group actions induced on corona algebras from a given $σ$-unital $\mathrm{C}^{\ast}$-dynamical system over a locally compact group $G$. We prove that such actions satisfy the so-called dynamical folding property, which generalizes a fundamental property observed for corona algebras in works

Yushi Hamaguchi

In this paper, we consider optimal control problems of stochastic Volterra equations (SVEs) with singular kernels, where the control domain is not necessarily convex. We establish a global maximum principle by means of the spike variation technique. To do so, we first show a Taylor type expansion of the controlled SVE with respect to the spike variation, whe

Andreas Jungherr, Adrian Rauchfleisch

Digital deliberation has expanded democratic participation, yet challenges remain. This includes processing information at scale, moderating discussions, fact-checking, or attracting people to participate. Recent advances in artificial intelligence (AI) offer potential solutions, but public perceptions of AI's role in deliberation remain underexplored. Beyon

Siyuan Song, Jennifer Hu, Kyle Mahowald

There has been recent interest in whether large language models (LLMs) can introspect about their own internal states. Such abilities would make LLMs more interpretable, and also validate the use of standard introspective methods in linguistics to evaluate grammatical knowledge in models (e.g., asking "Is this sentence grammatical?"). We systematically inves

Maxim Lisnic, Vidya Setlur, Nicole Sultanum

Text in dashboards plays multiple critical roles, including providing context, offering insights, guiding interactions, and summarizing key information. Despite its importance, most dashboarding tools focus on visualizations and offer limited support for text authoring. To address this gap, we developed Plume, a system to help authors craft effective dashboa

Chengmeng Li, Junjie Wen, Yan Peng, Yaxin Peng

Vision-Language-Action (VLA) models excel at robotic tasks by leveraging large-scale 2D vision-language pretraining, but their reliance on RGB images limits spatial reasoning critical for real-world interaction. Retraining these models with 3D data is computationally prohibitive, while discarding existing 2D datasets wastes valuable resources. To bridge this

Hari Shankar, Vedanta S P, Tejas Cavale, Ponnurangam Kumaraguru

Large Language Models (LLMs) are capable of generating opinions and propagating bias unknowingly, originating from unrepresentative and non-diverse data collection. Prior research has analysed these opinions with respect to the West, particularly the United States. However, insights thus produced may not be generalized in non-Western populations. With the wi

Mojtaba Vaezi, Xinliang Zhang

Non-orthogonal multiple access (NOMA) has gained significant attention as a potential next-generation multiple access technique. However, its implementation with finite-alphabet inputs faces challenges. Particularly, due to inter-user interference, superimposed constellations may have overlapping symbols leading to high bit error rates when successive interf

Jie Hu, Shizun Wang, Xinchao Wang

Recent advances in 2D-to-3D perception have enabled the recovery of 3D scene semantics from unposed images. However, prevailing methods often suffer from limited generalization, reliance on per-scene optimization, and semantic inconsistencies across viewpoints. To address these limitations, we introduce PE3R, a tuning-free framework for efficient and general

Soumya Banerjee, Vinay Kumar Verma

Active learning aims to select optimal samples for labeling, minimizing annotation costs. This paper introduces a unified representation learning framework tailored for active learning with task awareness. It integrates diverse sources, comprising reconstruction, adversarial, self-supervised, knowledge-distillation, and classification losses into a unified V

Qiongxiu Li, Wenrui Yu, Yufei Xia, Jun Pang

Federated Learning (FL) enables collaborative learning without directly sharing individual's raw data. FL can be implemented in either a centralized (server-based) or decentralized (peer-to-peer) manner. In this survey, we present a novel perspective: the fundamental difference between centralized FL (CFL) and decentralized FL (DFL) is not merely the network

Xiao Yan, Yi Ding

Recent advancements in large language models (LLMs) have prompted interest in deploying these models on mobile devices to enable new applications without relying on cloud connectivity. However, the efficiency constraints of deploying LLMs on resource-limited devices present significant challenges. In this paper, we conduct a comprehensive measurement study t

Seungjae Baek, Brady Moon, Seungchan Kim, Muqing Cao

Autonomous exploration in unknown environments requires estimating the information gain of an action to guide planning decisions. While prior approaches often compute information gain at discrete waypoints, pathwise integration offers a more comprehensive estimation but is often computationally challenging or infeasible and prone to overestimation. In this w

Shiu-hong Kao, Yu-Wing Tai, Chi-Keung Tang

Reasoning segmentation is a challenging vision-language task that aims to output the segmentation mask with respect to a complex, implicit, and even non-visual query text. Previous works incorporated multimodal Large Language Models (MLLMs) with segmentation models to approach the difficult problem. However, their segmentation quality often falls short in co

Christiaan Boerkamp, Akhil John Thomas

Medical image segmentation, particularly tumor segmentation, is a critical task in medical imaging, with U-Net being a widely adopted convolutional neural network (CNN) architecture for this purpose. However, U-Net's high computational and memory requirements pose challenges for deployment on resource-constrained devices such as wearable medical systems. Thi

A. Borrero, A. Díaz-Acosta, S. Blazquez, I. M. Zerón

In this work, the cryoscopic decrease effect, as a function of the NaCl concentration, on the carbon dioxide (CO$_2$) hydrate dissociation line conditions has been determined through molecular dynamic simulations. In particular, we have determined the three-phase (solid hydrate-aqueous phase-liquid CO$_2$) coexistence temperature at 100, 400, and 1000 at sev

Qiongxiu Li, Xiaoyu Luo, Yiyi Chen, Johannes Bjerva

The role of memorization in machine learning (ML) has garnered significant attention, particularly as modern models are empirically observed to memorize fragments of training data. Previous theoretical analyses, such as Feldman's seminal work, attribute memorization to the prevalence of long-tail distributions in training data, proving it unavoidable for sam

Junia S. Solomon, Nada Mrkyvkova, Vojtech Kliner, Tatiana Soto-Montero

Two-dimensional (2D) Ruddlesden-Popper (RP) Metal Halides present unique and tunable properties. However, direct and oriented synthesis is challenging due to low formation energies that lead to rapid, uncontrolled growth during solution-based processing. Here, we report the solvent-free growth of oriented and n = 1 2D $(\mbox{PEA})_2\mbox{PbI}_4$ RP films by

Calvin Yeung, Tomohiro Suzuki, Ryota Tanaka, Zhuoer Yin

Human pose estimation is a critical task in computer vision and sports biomechanics, with applications spanning sports science, rehabilitation, and biomechanical research. While significant progress has been made in monocular 3D pose estimation, current datasets often fail to capture the complex, high-acceleration movements typical of competitive sports. In

Saeed S. Jahromi, Yasir Iqbal

We study the ground-state phase diagram of the spin-$1/2$ antiferromagnetic Heisenberg model on the square-kagome lattice using infinite projected entangled-pair states (iPEPS). By systematically varying the ratio of exchange couplings on triangular and square plaquettes, we establish a complete quantum phase diagram in the thermodynamic limit. In the interm

Vladimir Markov

We examine the problem of optimal portfolio allocation within the framework of utility theory. We apply exponential utility to derive the optimal diversification strategy and logarithmic utility to determine the optimal leverage. We enhance existing methodologies by incorporating compound probability distributions to model the effects of both statistical and

Madhav Rijal, Rashik Shrestha, Trevor Smith, Yu Gu

This study presents a methodology to safely manipulate branches to aid various agricultural tasks. Humans in a real agricultural environment often manipulate branches to perform agricultural tasks effectively, but current agricultural robots lack this capability. This proposed strategy to manipulate branches can aid in different precision agriculture tasks,