May 2025 arXiv papers — page 42

Yang Shi, Huanqian Wang, Wulin Xie, Huanyao Zhang

Multimodal Large Language Models (MLLMs) have achieved considerable accuracy in Optical Character Recognition (OCR) from static images. However, their efficacy in video OCR is significantly diminished due to factors such as motion blur, temporal variations, and visual effects inherent in video content. To provide clearer guidance for training practical MLLMs

Christos Fragkathoulas, Evaggelia Pitoura

Counterfactual explanations (CFEs) are a popular approach for interpreting machine learning predictions by identifying minimal feature changes that alter model outputs. However, in real-world settings, users often refine feasibility constraints over time, requiring counterfactual generation to adapt dynamically. Existing methods fail to support such iterativ

Allaa Boutaleb, Bernd Amann, Hubert Naacke, Rafael Angarita

Recent table representation learning and data discovery methods tackle table union search (TUS) within data lakes, which involves identifying tables that can be unioned with a given query table to enrich its content. These methods are commonly evaluated using benchmarks that aim to assess semantic understanding in real-world TUS tasks. However, our analysis

Michael R. Blanton, Joleen K. Carlberg, Tom Dwelly, Ilija Medan

We present an algorithmic method for efficiently planning a long-term, large-scale multi-object spectroscopy program. The Sloan Digital Sky Survey V (SDSS-V) Focal Plane System performs multi-object spectroscopy using 500 robotic positioners to place fibers feeding optical and infrared spectrographs across a wide field. SDSS-V uses this system to observe tar

Md Mamun Miah, Ryuhei Wakida, Masato Kimura

Phase field models have emerged as a powerful and flexible framework for simulating complex interface-driven phenomena across a wide range of scientific and engineering applications. In fracture mechanics, the phase field approach--formulated as a gradient flow of the Griffith fracture energy with Ambrosio-Tortorelli regularization--has gained significant at

Jiakang Yuan, Tianshuo Peng, Yilei Jiang, Yiting Lu

Logical reasoning is a fundamental aspect of human intelligence and an essential capability for multimodal large language models (MLLMs). Despite the significant advancement in multimodal reasoning, existing benchmarks fail to comprehensively evaluate their reasoning abilities due to the lack of explicit categorization for logical reasoning types and an uncl

Kristopher Tapp, Todd Proebsting, Alec Ramsay

Ensemble analysis has become central to redistricting litigation, but parameter effects remain understudied. We analyze 315 ReCom ensembles across the three legislative chambers in 7 states, systematically varying the population tolerance, county preservation strength, and algorithm variant. To validate convergence, we introduce new methods to approximate ef

Guangyuan Li, Siming Zheng, Hao Zhang, Jinwei Chen

Video Virtual Try-On (VVT) aims to synthesize garments that appear natural across consecutive video frames, capturing both their dynamics and interactions with human motion. Despite recent progress, existing VVT methods still suffer from inadequate garment fidelity and limited spatiotemporal consistency. The reasons are: (1) under-exploitation of garment inf

Yuxin Zhu, Yuting Guo, Noah Marchuck, Abeed Sarker

Despite rapid advances in large language models (LLMs), their integration with traditional supervised machine learning (ML) techniques that have proven applicability to medical data remains underexplored. This is particularly true for psychiatric applications, where narrative data often exhibit nuanced linguistic and contextual complexity, and can benefit fr

Martin Škoudlil, Michal Sojka, Zdeněk Hanzálek

The increasing popularity of the Rust programming language in building robotic applications using the Robot Operating System (ROS 2) raises questions about its real-time execution capabilities, particularly when employing asynchronous programming. Existing real-time scheduling and response-time analysis techniques for ROS 2 focus on applications written in C

R. Spencer Hallyburton, Miroslav Pajic

Many state-of-the-art AI models deployed in cyber-physical systems (CPS), while highly accurate, are simply pattern-matchers.~With limited security guarantees, there are concerns for their reliability in safety-critical and contested domains. To advance assured AI, we advocate for a paradigm shift that imbues data-driven perception models with symbolic struc

Leonard Papenmeier, Luigi Nardi

We present Bencher, a modular benchmarking framework for black-box optimization that fundamentally decouples benchmark execution from optimization logic. Unlike prior suites that focus on combining many benchmarks in a single project, Bencher introduces a clean abstraction boundary: each benchmark is isolated in its own virtual Python environment and accesse

Sheng Zhao, Ya-Long Ren, Xin-Lei Hei, Xue-Feng Pan

Magnon blockade is a physical mechanism for the preparation of a single-magnon source, which has important applications in quantum information processing. Here we propose a scheme for generating an optimal magnon blockade in the spin-magnon quantum system. By introducing frequency detuning between the magnon and the spin qubit of the NV center, the conventio

Biao Zhang, Peter Wonka

Function fitting/approximation plays a fundamental role in computer graphics and other engineering applications. While recent advances have explored neural networks to address this task, these methods often rely on architectures with many parameters, limiting their practical applicability. In contrast, we pursue high-quality function approximation using para

Hao Li, He Cao, Bin Feng, Yanjun Shao

While large language models (LLMs) with Chain-of-Thought (CoT) reasoning excel in mathematics and coding, their potential for systematic reasoning in chemistry, a domain demanding rigorous structural analysis for real-world tasks like drug design and reaction engineering, remains untapped. Current benchmarks focus on simple knowledge retrieval, neglecting st

Ihab Bendidi, Yassir El Mesbahi, Alisandra K. Denton, Karush Suri

Understanding cellular responses to stimuli is crucial for biological discovery and drug development. Transcriptomics provides interpretable, gene-level insights, while microscopy imaging offers rich predictive features but is harder to interpret. Weakly paired datasets, where samples share biological states, enable multimodal learning but are scarce, limiti

Enam Ahmed Taufik, Antara Firoz Parsa, Seraj Al Mahmud Mostafa

Enhancing plant disease detection from leaf imagery remains a persistent challenge due to scarce labeled data and complex contextual factors. We introduce a transformative two-stage methodology, Mid Point Normalization (MPN) for intelligent image preprocessing, coupled with sophisticated attention mechanisms that dynamically recalibrate feature representatio

Jesujoba O. Alabi, Michael A. Hedderich, David Ifeoluwa Adelani, Dietrich Klakow

With over 2,000 languages and potentially millions of speakers, Africa represents one of the richest linguistic regions in the world. Yet, this diversity is scarcely reflected in state-of-the-art natural language processing (NLP) systems and large language models (LLMs), which predominantly support a narrow set of high-resource languages. This exclusion not

Qiao Yue, Zhaoyi Xu, Meirong Tang

This research delves into the optical characteristics of stationary, spherically symmetric black holes. These black holes follow the Konoplya-Zhidenko deformation rule in arbitrary gravity theories. This research finds that the effects of \(a_2\) and \(b_2\) on photon orbital dynamics exhibit observational degeneracy, while \(\varepsilon\) significantly gove

Thomas King, Sergio Vinciguerra

Understanding how faults nucleate and grow is a critical problem in earthquake science and hazard assessment. This study examines fault development in Alzo granite under triaxial pressures ranging from 5 to 40 MPa by applying a Time Delay Neural Network (TDNN) to multi-parameter acoustic emission (AE) data. The TDNN integrates waveform-derived attributes, in

Y. L. Wang, Y. K. Hsiao

Using the topological-diagram approach based on $SU(3)$ flavor symmetry, we investigate two-body $\Lambda_c^+\to {\bf B}S$ decays, where ${\bf B}$ denotes the final-state baryon and $S$ refers to a light scalar meson, such as $f_0/f_0(980)$, $a_0/a_0(980)$, $\sigma_0/f_0(500)$, or $\kappa/K_0^*(700)$. Our analysis indicates that interpreting the light scalar

Aihua Fan, Shilei Fan

This paper establishes two fundamental results on the existence of exponential Riesz basis in non-Archimedean locally compact Abelian groups: the existence of Riesz basis of exponentials for all finite unions of balls and the non-existence of such basis for some bounded sets.

Zenghao Zheng, Lianping Yang, Hegui Zhu, Mingrui Ye

Transformer-based 3D human pose estimation methods suffer from high computational costs due to the quadratic complexity of self-attention with respect to sequence length. Additionally, pose sequences often contain significant redundancy between frames. However, recent methods typically fail to improve model capacity while effectively eliminating sequence red

Lin Lin

While dissipation has traditionally been viewed as an obstacle to quantum coherence, it is increasingly recognized as a powerful computational resource. Dissipative protocols can prepare complex many-body quantum states by leveraging engineered system-environment interactions. This essay focuses on a class of algorithms that utilize algorithmically construct

L. A. Busch, J. E. Pineda, O. Sipilä, D. M. Segura-Cox

The 12C/13C isotope ratio has been derived towards numerous cold clouds (20-50 K) and a couple protoplanetary disks and exoplanet atmospheres. However, direct measurements of this ratio in the warm gas (>100 K) around young low-mass protostars remain scarce, but are required to study its evolution during star and planet formation. We derived 12C/13C ratios f

Wesley Pegden, Francesca Yu

In classical Maker-Breaker games on graphs, Maker and Breaker take turns claiming edges; Maker's goal is to claim all of some structure (e.g., a spanning tree, Hamilton cycle, etc.), while Breaker aims to stop her. The standard question considered is how powerful a Breaker Maker can defeat; i.e., for the $(1:b)$-biased game where Breaker takes $b$ edges per

Shirin Chenarani, Mojtaba Mohammadi Najafabadi

Axion-like particles (ALPs) are pseudo Nambu-Goldstone bosons associated with spontaneously broken global symmetries incorporated in the Standard Model (SM) Lagrangian in many models beyond the SM. The existence of a light ALP is plausible due to the long-standing problems that the SM has not been able to address, such as the dark matter (DM) problem and the

Louis Jalouzot, Alexis Thual, Yair Lakretz, Christophe Pallier

We investigate optimal strategies for decoding perceived natural speech from fMRI data acquired from a limited number of participants. Leveraging Lebel et al. (2023)'s dataset of 8 participants, we first demonstrate the effectiveness of training deep neural networks to predict LLM-derived text representations from fMRI activity. Then, in this data regime, we

Libo Xie, Liang-Bi Wu, Zong-Kuan Guo

We investigate the stability of QNM spectra and greybody factors in the Schwarzschild black hole by approximating the Regge-Wheeler potential with a piecewise parabolic form and treating the deviation as a perturbation. We find that QNM spectra are sensitive to small perturbations, while greybody factors remain stable. This piecewise parabolic approximated p

Bartosz Błasiak, Dominik Brey, Rocco Martinazzo, Irene Burghardt

Thermofield dynamics (TFD) is a powerful framework to account for thermal effects in a wavefunction setting, and has been extensively used in physics and quantum optics. TFD relies on a duplicated state space and creates a correlated two-mode thermal state via a Bogoliubov transformation acting on the vacuum state. However, a very useful variant of TFD uses

Andrea Pedrotti, Giulia Rambelli, Caterina Villani, Marianna Bolognesi

People can categorize the same entity at multiple taxonomic levels, such as basic (bear), superordinate (animal), and subordinate (grizzly bear). While prior research has focused on basic-level categories, this study is the first attempt to examine the organization of categories by analyzing exemplars produced at the subordinate level. We present a new Itali

Servet Ozdemir, Matthew Rogers, Jaka Strohsack, Hari Babu Vasili

Information can be stored in magnetic materials by encoding with the direction of the magnetic moment of elements. A figure of merit for these systems is the energy needed to change the information rewrite the storage by changing the magnetic moment. Organic molecules offer a playground to manipulate spin order, with metallo molecular interfaces being a prom

Alexa Gopaulsingh, Zalán Molnár, Amitayu Banerjee

A distinguishing coloring of a graph is a vertex coloring such that only the identity automorphism of the graph preserves the coloring. A 2-distinguishable graph is a graph which can be distinguished using 2 colors. The cost $\rho(G)$ of a 2-distinguishable graph is the smallest size of a color set of a distinguishing coloring of $G$. The determining number

Emanuele La Malfa, Gabriele La Malfa, Samuele Marro, Jie M. Zhang

Recent interest in Multi-Agent Systems of Large Language Models (MAS LLMs) has led to an increase in frameworks leveraging multiple LLMs to tackle complex tasks. However, much of this literature appropriates the terminology of MAS without engaging with its foundational principles. In this position paper, we highlight critical discrepancies between MAS theory

Yifei Liu, Li Lyna Zhang, Yi Zhu, Bingcheng Dong

Advancing code reasoning in large language models (LLMs) is fundamentally limited by the scarcity of high-difficulty datasets, especially those with verifiable input-output test cases necessary for rigorous solution validation at scale. We introduce rStar-Coder, which significantly improves LLM code reasoning capabilities by constructing a large-scale, verif

Nicolas Eschenbaum, Nicolas Greber

We analyze the vulnerability of decentralized autonomous organizations (DAOs) to speculative exploitation via their redemption mechanisms. Studying a game-theoretic model of repeated auctions for governance shares with speculators, we characterize the conditions under which -- in equilibrium -- an exploitative exit is guaranteed to occur, occurs in expectati

A. Belayadi, C. I. Osuala, I. Assi, A. Naif

Graphene/hexagonal boron nitride (hBN) heterostructures represent a promising class of metal-insulator-semiconductor systems widely explored for multifunctional digital device applications. In this work, we demonstrate that graphene, when influenced by carrier-dependent trapping in the hBN spacer triggered by a localized potential from Kelvin probe force mic

Agathe Bouis, Ruaridh A. Clark, Malcolm Macdonald

A network's community structure commonly impacts its functions. For instance, networks seeking synchronisation will see this process follow the topology's hierarchical and community structuring. Herein, the interplay of network adjacency and Laplacian eigenspectra is shown to uncover hierarchical influence and community structure. Ring networks embedded with

Bennet Karetta, Xanthe H. Verbeek, Rodrigo Jaeschke-Ubiergo, Libor Šmejkal

The possibility of a strain-induced transformation from $g$-wave to $d$-wave altermagnetism was recently recently proposed for MnTe using a $k\cdot p$ perturbative model. In this work, we demonstrate such a transition in CrSb for a wider array of strains, using a combination of a minimal model and first-principles calculations. Starting from a symmetry persp

Morgan H. Lynch

In this manuscript we examine the high energy channeling radiation data sets from the CERN-NA63 experiment using ultra relativistic synchrotron emission. To incorporate recoil, we examine the standard quasi-classical formalism as well as develop a formalism which includes the Unruh effect by utilizing a hyperbolic recoil acceleration, based on conservation o

Yeshwanth Venkatesha, Souvik Kundu, Priyadarshini Panda

Large Language Models (LLMs) enable various applications on edge devices such as smartphones, wearables, and embodied robots. However, their deployment often depends on expensive cloud-based APIs, creating high operational costs, which limit access for smaller organizations and raise sustainability concerns. Certain LLMs can be deployed on-device, offering a

Saman Marandi, Yu-Shu Hu, Mohammad Modarres

In this paper, we present a novel diagnostic framework that integrates Knowledge Graphs (KGs) and Large Language Models (LLMs) to support system diagnostics in high-reliability systems such as nuclear power plants. Traditional diagnostic modeling struggles when systems become too complex, making functional modeling a more attractive approach. Our approach in

Colin Cooper, Alan Frieze

Let $G_{n,p}^{[\kappa]}$ denote the space of $n$-vertex edge coloured graphs, where each edge occurs independently with probability $p$. The colour of each existing edge is chosen independently and uniformly at random from the set $[\kappa]$. We consider the threshold for the existence of rainbow colored copies of a spanning subgraph $H$. We provide lower bo

Nurbek Tastan, Stefanos Laskaridis, Martin Takac, Karthik Nandakumar

Large pre-trained models are commonly adapted to downstream tasks using parameter-efficient fine-tuning methods such as Low-Rank Adaptation (LoRA), which injects small trainable low-rank matrices instead of updating all weights. While LoRA dramatically reduces trainable parameters with little overhead, it can still underperform full fine-tuning in accuracy a

Farshad Noravesh, Reza Haffari, Layki Soon, Arghya Pal

Graph Attention Networks (GATs) have emerged as powerful models for learning expressive representations from such data by adaptively weighting neighboring nodes through attention mechanisms. However, most existing approaches primarily rely on node attributes and direct neighborhood connections, often overlooking rich structural patterns that capture higher-o

E. Cancès, A. Kirsch, S. Perrin--Roussel

In a previous contribution (E. Canc\`es, A. Kirsch and S. Perrin--Roussel, arXiv:2406.03384), we have proven the existence of a solution to the Dynamical Mean-Field Theory (DMFT) equations under the Iterated Perturbation Theory (IPT-DMFT) approximation. In view of numerical simulations, these equations need to be discretized. In this article, we are interest

Ya-Ting Yang, Quanyan Zhu

Agentic AI, often powered by large language models (LLMs), is becoming increasingly popular and adopted to support autonomous reasoning, decision-making, and task execution across various domains. While agentic AI holds great promise, its deployment as services for easy access raises critical challenges in pricing, due to high infrastructure and computation

Dipam Goswami, Liying Wang, Bartłomiej Twardowski, Joost van de Weijer

Text embedding models enable semantic search, powering several NLP applications like Retrieval Augmented Generation by efficient information retrieval (IR). However, text embedding models are commonly studied in scenarios where the training data is static, thus limiting its applications to dynamic scenarios where new training data emerges over time. IR metho

Matthias Dietl, Marco Valentini, Fabian Anmasser, Alexander Zesar

Ion traps are a promising architecture to host a future quantum computer. Several challenges, such as signal-routing, power dissipation, and fabrication quality need to be overcome to scale ion trap devices to hundreds of ions. Currently, ion traps are often fabricated on silicon substrates which result in high power dissipation. Substrates that lead to lowe

Yu-Ji Shi, Wei Wang, Ji Xu

We conduct an SU(3) analysis of $B\to PP$ decays based on reduced matrix elements (RMEs), with $P$ being a light pseudoscalar meson excluding $\eta^{(\prime)}$. We show that a complete basis for the $B\to PP$ decays consists of ten RMEs, where the three RMEs arise from the electroweak penguin operators $O_{7,8}$. In the Standard Model, the relevant Wilson co

Timur Akhtyamov, Mohamad Al Mdfaa, Javier Antonio Ramirez Benavides, Arthur Nigmatzyanov

Data-driven navigation algorithms are critically dependent on large-scale, high-quality real-world data collection for successful training and robust performance in realistic and uncontrolled conditions. To enhance the growing family of navigation-related real-world datasets, we introduce EgoWalk - a dataset of 50 hours of human navigation in a diverse set o

Yue Zhang, Zhiliang Tian, Shicheng Zhou, Haiyang Wang

Legal Judgment Prediction (LJP) is a pivotal task in legal AI. Existing semantic-enhanced LJP models integrate judicial precedents and legal knowledge for high performance. But they neglect legal reasoning logic, a critical component of legal judgments requiring rigorous logical analysis. Although some approaches utilize legal reasoning logic for high-qualit

Rafael Acuna, Aldie Alejandro, Robert Leung

Dynasties have long dominated Philippine politics. Despite the theoretical consensus that dynastic rule erodes democratic accountability, there is limited empirical evidence establishing dynasties' true impact on development. A key challenge has been developing robust metrics for characterizing dynasties that facilitate meaningful comparisons across geograph

Shaoqing Zhang, Kehai Chen, Zhuosheng Zhang, Rumei Li

Recent advancements in vision-language models have increased interest in Device-Control Agents (DC agents) for managing graphical user interfaces (GUIs). With the growing complexity and integration of such agents into various applications, effective evaluation methods have become crucial. The current evaluation method for DC agents primarily focuses on the i

Lukas Bauer, Ekaterina Kazak

This paper proposes a Conditional Method Confidence Set (CMCS) which allows to select the best subset of forecasting methods with equal predictive ability conditional on a specific economic regime. The test resembles the Model Confidence Set by Hansen et al. (2011) and is adapted for conditional forecast evaluation. We show the asymptotic validity of the pro

Yao Huang, Yitong Sun, Shouwei Ruan, Yichi Zhang

Large Language Models (LLMs), despite advanced general capabilities, still suffer from numerous safety risks, especially jailbreak attacks that bypass safety protocols. Understanding these vulnerabilities through black-box jailbreak attacks, which better reflect real-world scenarios, offers critical insights into model robustness. While existing methods have

David Winkelmann, Christian Deutscher

We use the fertile ground of betting markets to study the anticipation of major news in financial markets. While there is a considerable body of literature on the accuracy and efficiency of betting markets after important in-match events, there are no studies dealing with the anticipation of such events. This paper tracks bookmaker odds and betting stakes to

Chen Liu, Hengyu Tang, Zhixiao Yang, Ke Zhou

In the age of digital finance, detecting fraudulent transactions and money laundering is critical for financial institutions. This paper presents a scalable and efficient solution using Big Data tools and machine learning models. We utilize realtime data streaming platforms like Apache Kafka and Flink, distributed processing frameworks such as Apache Spark,

Léo Portales, Edouard Pauwels, Elsa Cazelles

Computational implementation of optimal transport barycenters for a set of target probability measures requires a form of approximation, a widespread solution being empirical approximation of measures. We provide an $O(\sqrt{N/n})$ statistical generalization bounds for the empirical sparse optimal transport barycenters problem, where $N$ is the maximum cardi

J. Le Bourlot, E. Roueff, S. R. Federman, A. M. Ritchey

Context. Recent spectroscopic measurements have revealed absorption from higher rotational levels in C$_2$ than previous observations. These improvements are accompanied by the availability of updated radiative and collisional data. Aims. We revisit the density and radiation field intensity diagnostics provided by the observations of many rotational levels o

Eugenia O'Reilly-Regueiro, Octavio B. Zapata-Fonseca

This paper was inspired by a paper by Blokhuis and Brouwer [Designs, Codes and Cryptography 65, 2012] in which a definition of a graph on the flags of a biplane is given, and they prove that the graph corresponding to the unique $(11,5,2)$-biplane is determined by its spectrum. It is also inspired by the different definition of flag-graph seen in the context

Mathieu Dumberry

The mantle-inner core gravitational (MICG) mode is the free mode axial oscillation between the mantle and inner core sustained by the gravitational torque between their degree 2 order 2 density structures. Here, we investigate how the MICG mode is affected by oscillations of cylindrical surfaces in the fluid outer core in the form of Alfv\'en waves. The latt

Wenqin Wang, Xinyue Ma, Zhangliang Chen, Cong Yu

Transit timing variation (TTV) is a useful tool for studying the orbital properties of transiting objects. However, few TTV studies have been done on transiting brown dwarfs (BDs) around solar-type stars. Here we study the long-term TTV of a population of close BD companions around solar-type stars using TESS data. We use the measured orbital period change r

Eva Gmelich Meijling, Roberto Del Prete, Arnoud Visser

Accurate wetland land-cover classification is essential for environmental monitoring, biodiversity assessment, and sustainable ecosystem management. However, the scarcity of annotated data, especially for high-resolution satellite imagery, poses a significant challenge for supervised learning approaches. To tackle this issue, this study presents a methodolog

David Norrbo

We study the interchange of essential norm and integration of certain families of weighted composition operators acting on the standard weighted Bergman spaces $A^p_\alpha$, where $p>1$ and $\alpha\geq 0$. To be more precise, we give a sufficient condition for $ \|\int u_tC_{\phi_t}\, dt\|_e = \int \| u_tC_{\phi_t}\|_e \, dt $ to hold in terms of geometric p

Nguyen N. Hung, J. Miquel Martínez, Gabriel Navarro

We study the sum of the squares of the irreducible character degrees not divisible by some prime $p$, and its relationship with the the corresponding quantity in a $p$-Sylow normalizer. This leads to study a recent conjecture by E. Giannelli, which we prove for $p=2$ and in some other cases.

Eve Le Guillou, Pierre Fortin, Julien Tierny

The persistence diagram, which describes the topological features of a dataset, is a key descriptor in Topological Data Analysis. The "Discrete Morse Sandwich" (DMS) method has been reported to be the most efficient algorithm for computing persistence diagrams of 3D scalar fields on a single node, using shared-memory parallelism. In this work, we extend DMS

Ilker Kesen, Jonas F. Lotz, Ingo Ziegler, Phillip Rust

Pixel language models operate directly on images of rendered text, eliminating the need for a fixed vocabulary. While these models have demonstrated strong capabilities for downstream cross-lingual transfer, multilingual pretraining remains underexplored. We introduce PIXEL-M4, a model pretrained on four visually and linguistically diverse languages: English

Peter Scherbak, Wenbin Lu, Jim Fuller

High rates of stable mass transfer likely occur for some binary star systems, but the resulting flow of mass and angular momentum (AM) is unclear. We perform hydrodynamical simulations of a polytropic donor star and a point mass secondary to determine the mass, AM, and velocity of gas that escapes the system, and the dependence on binary parameters such as m

Avihay Cohen

Modern software supply chains face an increasing threat from malicious code hidden in trusted components such as browser extensions, IDE extensions, and open-source packages. This paper introduces JavaSith, a novel client-side framework for analyzing potentially malicious extensions in web browsers, Visual Studio Code (VSCode), and Node's NPM packages. JavaS

M. Akin Yilmaz, Ahmet Bilican, A. Murat Tekalp

Balancing reconstruction quality versus model efficiency remains a critical challenge in lightweight single image super-resolution (SISR). Despite the prevalence of attention mechanisms in recent state-of-the-art SISR approaches that primarily emphasize or suppress feature maps, alternative architectural paradigms warrant further exploration. This paper intr

Ethan Marx, William Benoit, Trevor Blodgett, Deep Chatterjee

We conduct a search for stellar-mass binary black hole mergers in gravitational-wave data collected by the LIGO detectors during the LIGO-Virgo-KAGRA (LVK) third observing run (O3). Our search uses a machine learning (ML) based method, Aframe, an alternative to traditional matched filtering search techniques. The O3 observing run has been analyzed by the LVK

Heinrich Steigerwald

The dark matter (DM) conundrum is one of the most intriguing due to its resistance in direct detection experiments. In recent years, attempts to identify non-gravitational signatures as the result of DM traversing or accumulating within stars have attracted a lot of attention. These calculations are usually evaluated at the order-of-magnitude level for stell

Chunyi Ma, Jiajie Xu, Jianhua Yang, Mustafa A. Kishk

To achieve the Internet of Things (IoT) vision,Mobile Edge Computing (MEC) is a promising technology aimed at providing low-latency computing services to user equipment (UE). However, terrestrial MEC network struggles to provide service to UEs in remote and maritime region. Low Earth Orbit (LEO) satellite networks have the potential to overcome geographical

Changguanng Wu, Jiangxin Dong, Chengjian Li, Jinhui Tang

We present Plenodium (plenoptic medium), an effective and efficient 3D representation framework capable of jointly modeling both objects and participating media. In contrast to existing medium representations that rely solely on view-dependent modeling, our novel plenoptic medium representation incorporates both directional and positional information through

Giacomo Canevari, Van Phu Cuong Le, Ramon Oliver-Bonafoux, Giandomenico Orlandi

We prove a ${\Gamma}$-convergence result for the $p$-Dirichlet energy functional defined on maps from a smooth bounded domain $\Omega \subseteq \mathbb{R}^{n+k}$ to $\mathscr{N}$, a $(k-2)$-connected and smooth closed Riemannian manifold with Abelian fundamental group, where $n$ and $k$ are integers, $n \geq 0$, $k \geq 2$. We focus on the regime $p \to~k^-$

Heinrich Steigerwald

We set up and perform collision rate simulations between dark matter in the form of asteroid-mass primordial black holes (PBHs) and white dwarf stars. These encounters trigger prompt detonations and could be the key to solving the ignition mystery of type Ia supernovae. Our framework is flexible enough to cover the full range of progenitor white dwarf masses

Mauro D'Arcangelo, Younes Javanmard, Natalie Pearson

In this work, we present a quantum Markov chain algorithm for many-body systems that utilizes a special phase of matter known as the Many-Body Localized (MBL) phase. We show how the properties of the MBL phase enable one to address the conditions for ergodicity and sampling from distributions of quantum states. We demonstrate how to exploit the thermalized-t

Jake Langham, Xiannan Meng, Jamie P. Webb, Chris G. Johnson

Depth-averaged systems of equations describing the motion of fluid-sediment mixtures have been widely adopted by scientists in pursuit of models that can predict the paths of dangerous overland flows of debris. As models have become increasingly sophisticated, many have been developed from a multi-phase perspective in which separate, but mutually coupled set

Yuki Miyamoto

Let $G$ be a locally compact group, $L(G)$ be its group von Neumann algebra equipped with the Plancherel weight $\varphi_G$. In this paper, we consider the following two questions. (1) When is the restriction of $\varphi_G$ to the subalgebra generated by a closed subgroup $H$ semifinite? If so, is it equal (up to a constant) to $\varphi_H$? (2) When is $L(G)

Yang Yang, Siming Zheng, Qirui Yang, Jinwei Chen

Diffusion models have recently emerged as powerful tools for camera simulation, enabling both geometric transformations and realistic optical effects. Among these, image-based bokeh rendering has shown promising results, but diffusion for video bokeh remains unexplored. Existing image-based methods are plagued by temporal flickering and inconsistent blur tra

Aalok Gangopadhyay, Prajwal Singh, Ashish Tiwari, Shanmuganathan Raman

Shadow art is an exciting form of sculptural art that produces captivating artistic effects through the 2D shadows cast by 3D shapes. Hand shadows, also known as shadow puppetry or shadowgraphy, involve creating various shapes and figures using your hands and fingers to cast meaningful shadows on a wall. In this work, we propose a differentiable rendering-ba

Ze Chen, Shaode Yu

Kolmogorov-Arnold Network (KAN) has attracted growing interest for its strong function approximation capability. In our previous work, KAN and its variants were explored in score regression for blind image quality assessment (BIQA). However, these models encounter challenges when processing high-dimensional features, leading to limited performance gains and

Mustafa Hajij, Lennart Bastian, Sarah Osentoski, Hardik Kabaria

We introduce copresheaf topological neural networks (CTNNs), a powerful unifying framework that encapsulates a wide spectrum of deep learning architectures, designed to operate on structured data, including images, point clouds, graphs, meshes, and topological manifolds. While deep learning has profoundly impacted domains ranging from digital assistants to a

Dosung Lee, Wonjun Oh, Boyoung Kim, Minyoung Kim

Multi-hop question answering (MHQA) involves reasoning across multiple documents to answer complex questions. Dense retrievers typically outperform sparse methods like BM25 by leveraging semantic embeddings; however, they require labeled query-document pairs for fine-tuning. This poses a significant challenge in MHQA due to the high variability of queries (r

Mohamed Benzaghta, Sahar Ammar, David López-Pérez, Basem Shihada

Mobility management in cellular networks faces increasing complexity due to network densification and heterogeneous user mobility characteristics. Traditional handover (HO) mechanisms, which rely on predefined parameters such as A3-offset and time-to-trigger (TTT), often fail to optimize mobility performance across varying speeds and deployment conditions. F

Kui Xie, Giovanni Romagnoli, Giordana Bucchioni, Alberto Bemporad

Accurate relative orbit determination is a significant challenge in modern space operations, particularly when relying only on angular measurements. The inherent observability limitations of this approach make initial state estimation difficult, directly impacting mission safety and performance. This work proposes a hybrid estimation and control strategy for

Konstantin Karandashev

A computationally efficient protocol for machine learning in chemical space using Boltzmann ensembles of conformers as input is proposed; the method is based on rewriting Kernel Ridge Regression expressions in terms of Structured Orthogonal Random Features, yielding physics-motivated trigonometric neural networks. To evaluate the method's utility for materia

Karl Henrik Fredly, Tor Ole B. Odden, Benjamin M. Zwickl

Computation has become an integral part of physics research. However, little is known about how students learn to productively use computation as a tool beyond the introductory level, especially as they transition into physics research. In this study, we apply the theory of physics computational literacy and the novice-expert framework to describe the develo

Zhaoqing Li, Haoning Xu, Zengrui Jin, Lingwei Meng

Model compression has become an emerging need as the sizes of modern speech systems rapidly increase. In this paper, we study model weight quantization, which directly reduces the memory footprint to accommodate computationally resource-constrained applications. We propose novel approaches to perform extremely low-bit (i.e., 2-bit and 1-bit) quantization of

Ya-Ting Yang, Quanyan Zhu

Cyber deception is an emerging proactive defense strategy to counter increasingly sophisticated attacks such as Advanced Persistent Threats (APTs) by misleading and distracting attackers from critical assets. However, since deception techniques incur costs and may lose effectiveness over time, defenders must strategically time and select them to adapt to the

Colm Kelleher, Frédéric Holweck

We present definitive violations of non-contextual hidden variable bounds in the latest generation of IBM noisy intermediate-scale quantum computers (NISQ). These violations are based on known tests for contextuality such as the Rio Negro inequality and pseudo-telepathic Mermin games. These are the first violations of the classical Mermin game on IBM NISQ co

Gunjan Balde, Soumyadeep Roy, Mainack Mondal, Niloy Ganguly

Large Language Models (LLMs) recently achieved great success in medical text summarization by simply using in-context learning. However, these recent efforts do not perform fine-grained evaluations under difficult settings where LLMs might fail. They typically report performance scores over the entire dataset. Through our benchmarking study, we show that LLM

Divya Nori, Anisha Parsan, Caroline Uhler, Wengong Jin

Protein binder design has been transformed by hallucination-based methods that optimize structure prediction confidence metrics, such as the interface predicted TM-score (ipTM), via backpropagation. However, these metrics do not reflect the statistical likelihood of a binder-target complex under the learned distribution and yield sparse gradients for optimiz

Yahui Chai, Yibin Guo, Stefan Kühn

Scattering processes are fundamental for understanding the structure of matter, yet simulating their real-time dynamics remains challenging for classical computers. Quantum computing and quantum-inspired methods offer a promising avenue for efficiently simulating such phenomena. In this work, we investigate meson scattering in a (1+1)-dimensional Z2 lattice

Zhaoqing Li, Haoning Xu, Xurong Xie, Zengrui Jin

This paper presents a novel memory-efficient model compression approach for Conformer ASR and speech foundation systems. Our approach features a unique "small-to-large" design. A compact "seed" model containing a few Conformer or Transformer blocks is trained and unfolded many times to emulate the performance of larger uncompressed models with different logi

Felix Chalumeau, Daniel Rajaonarivonivelomanantsoa, Ruan de Kock, Claude Formanek

Reinforcement learning (RL) systems have countless applications, from energy-grid management to protein design. However, such real-world scenarios are often extremely difficult, combinatorial in nature, and require complex coordination between multiple agents. This level of complexity can cause even state-of-the-art RL systems, trained until convergence, to

Tomoyasu Yokoyama, Kazuhide Ichikawa, Hisashi Naito

Crystal structures can be viewed as assemblies of space-filling polyhedra, which play a critical role in determining material properties such as ionic conductivity and dielectric constant. However, most conventional crystal structure prediction methods rely on random structure generation and do not explicitly incorporate polyhedral tiling, limiting their eff

Elif Unsal, Alessandro Pecchia, Alexander Croy, Gianaurelio Cuniberti

Graphynes, a class of two-dimensional carbon allotropes, exhibit exceptional electronic properties, similar to graphene, but with intrinsic band gaps, making them promising for semiconducting applications. The incorporation of acetylene linkages allows for systematic modulation of their properties. However, the theoretical characterization of graphynes remai

Jiawei Guo, Feifei Zhai, Pu Jian, Qianrun Wei

Current VLM-based VQA methods often process entire images, leading to excessive visual tokens that include redundant information irrelevant to the posed question. This abundance of unnecessary image details creates numerous visual tokens, drastically increasing memory and computational requirements in VLMs. To address this, we propose Contextual Region-Orien