October 2025 arXiv papers — page 4

Forough Mehralian, Ryan Shar, James R. Rae, Alireza Hashemi

As large language models become increasingly capable of generating code, evaluating their performance remains a complex and evolving challenge. Existing benchmarks primarily focus on functional correctness, overlooking the diversity of real-world coding tasks and developer expectations. To this end, we introduce a multi-language benchmark that evaluates LLM

Simon Schulz, Ivan Yuri Violo

In the present article we prove second-order and Lipschitz regularity for quasilinear elliptic equations in metric spaces endowed with a lower bound on the Ricci curvature. The estimates we obtain are quantitative and cover a large class of elliptic equations with polynomial growth. As a particular case we settle the Lipschitz regularity of $p$-harmonic func

Volodymyr Vasylkovskyi, Andrey Evlyukhin, Elena Schlein, Mykola Slipchenko

Perovskite materials are at the forefront of modern materials science due to their exceptional structural, electronic, and optical properties. The controlled fabrication of perovskite nanostructures is crucial for enhancing their performance, stability, and scalability, directly impacting their applications in next-generation devices such as solar cells, LED

Yuchen Jiao, Yuchen Zhou, Gen Li

Score-based diffusion models have achieved remarkable empirical success in generating high-quality samples from target data distributions. Among them, the Denoising Diffusion Probabilistic Model (DDPM) is one of the most widely used samplers, generating samples via estimated score functions. Despite its empirical success, a tight theoretical understanding of

Joseph F. Wild, Yihan Li, Keyue Liang, Aishwarya S. Gujarathi

We present evidence that the Einstein relation (ER) breaks down completely in pure water and dilute aqueous solutions under strong centrifugation fields at 40 oC. Isotopologues (e.g., H2O-18) and solutes migrate at a speed of only 5% of that predicted based on the ER. The ER is restored with the addition of solutes above a transition concentration (ct). We f

Nagisa Hiroshima, Mitsuru Kakizaki, Shuhei Ohzawa

We investigate the trilinear Higgs boson coupling derived from the functional forms of various extended Higgs potentials. In light of experimental constraints on Higgs boson couplings, we focus on extended Higgs models in which the trilinear Higgs boson coupling is predominantly determined by the Standard Model (SM) Higgs field. Such models include the nearl

Raffaele D'Avino, Gabriel Senno, Mir Alimuddin, Antonio Acín

Quantum information tasks are often analyzed under varying trust assumptions about the devices involved. The semi-device-independent (SDI) framework offers a balance between needed assumptions and experimental feasibility. In this work, we study the energy-constrained SDI scenario, where the only assumption in a prepare-and-measure setup is an upper bound on

Sushil Samuel Dinesh, Shinkyu Park

This paper presents a framework that leverages pre-trained foundation models for robotic manipulation without domain-specific training. The framework integrates off-the-shelf models, combining multimodal perception from foundation models with a general-purpose reasoning model capable of robust task sequencing. Scene graphs, dynamically maintained within the

Johannes Gloßner

In this article we provide a model-independent definition of the concept of lax $2$-functors from $(\infty,2)$-category theory and show that it agrees with the existing and widely used combinatorial model for those in terms of inert-cocartesian functors, which is utilized for example in the foundational work of Gaitsgory and Rozenblyum on Derived Algebraic G

Lino Gerlach, Elliott Kauffman, Liv Helen Våge, Isobel Ojalvo

The increasing data rates and complexity of detectors at the Large Hadron Collider (LHC) necessitate fast and efficient machine learning models, particularly for rapid selection of what data to store, known as triggering. Building on recent work in differentiable logic gates, we present a public implementation of a Convolutional Differentiable Logic Gate Neu

Inacio Vieira, Antonio Castaldo, James O'Doherty, Sheila Castilho

LLMs often require adaptation to domain-specific requirements, a process that can be expensive when relying solely on SFT. We present an empirical study on applying CPO to simulate a post-editing workflow for data-efficient domain adaptation. Our approach synthesizes preference pairs by treating the base model's own raw output as the 'rejected' translation a

Redouane Douaifia, Salem Abdelmalek, Mokhtar Kirane

We investigate a class of three-component reaction-diffusion systems subject to mass control and a newly introduced structural assumption, referred to as linear intermediate weighted sum condition. Under these hypotheses, we establish the global existence of classical solutions in arbitrary spatial dimensions and wide class of boundary conditions, even when

David Shi, Kevin Joo

x402 enables Hypertext Transfer Protocol (HTTP) services like application programming interfaces (APIs), data feeds, and inference providers to accept cryptocurrency payments for access. As agents increasingly consume these services, discovery becomes critical: which swap interface should an agent trust? Which data provider is the most reliable? We introduce

M. McMurdo, T. Van Doorsselaere, N. Magyar, L. Banovic

The coronal heating problem remains a fundamental challenge in solar physics. While AWSoM-type models (Alfv\'en Wave Solar Model) have proven highly successful in reproducing the large-scale structure of the solar corona, they inherently neglect contributions from additional wave modes that arise when the effects of transverse structuring is fully incorporat

Yinghao Luo, Lang Zhou, Amrish Jhingoer, Klaske Vliegenthart Jongbloed

In multilingual healthcare applications, the availability of domain-specific natural language processing(NLP) tools is limited, especially for low-resource languages. Although multilingual bidirectional encoder representations from transformers (BERT) offers a promising motivation to mitigate the language gap, the medical NLP tasks in low-resource languages

Michela Frigeri, Veronica Berrocal, Alessandra Guglielmi

Understanding the sources that contribute to fine particulate matter (PM$_{2.5}$) is of crucial importance for designing and implementing targeted air pollution mitigation strategies. Determining what factors contribute to a pollutant's concentration goes under the name of source apportionment and it is a problem long studied by atmospheric scientists and st

Nenad Antonić, Darko Mitrović, Tomislav Perić

We extend G\'erard's results on orthogonality of ${\rm L}^2_{\rm loc}$ sequences as a consequence of mutual singularity of corresponding H-measures (microlocal defect measures) to ${\rm L}^p$/${\rm L}^q$ sequences and newly introduced notion of orhogonality for H-distributions. We apply the result to a homogenisation problem for the heterogeneous Boltzmann e

Steven Cunnington, Matilde Barberi-Squarotti, José Luis Bernal, Stefano Camera

Mapping the integrated 21cm emission line from dark matter-tracing neutral hydrogen gas is the primary science goal for MeerKLASS (MeerKAT's Large Area Synoptic Survey). Prior to the arrival of MeerKAT, this intensity mapping technique had only been tested on a couple of pre-existing single-dish radio telescopes with a handful of observational hours with whi

I. H. Senevirathne, T. Powers, N. Raut

Plasma processing of superconducting radio frequency (SRF) cavities has been an active research effort at Jefferson Lab (JLab) since 2019, aimed at enhancing cavity performance by removing hydrocarbon contaminants and reducing field emission. In this experiment, processing using argon-oxygen and helium-oxygen gas mixtures to find minimum ignition power at di

Xue Xia, Randall Balestriero, Tao Zhang, Yixin Zhou

Historical maps are unique and valuable archives that document geographic features across different time periods. However, automated analysis of historical map images remains a significant challenge due to their wide stylistic variability and the scarcity of annotated training data. Constructing linked spatio-temporal datasets from historical map time series

David Rovira, Ivan Caño, Cibran Lopez, Alejandro Navarro-Güell

Quasi-one-dimensional (Q-1D) van der Waals chalcohalides have emerged as promising materials for advanced energy applications, combining tunable optoelectronic properties and composed by earth-abundant and non-toxic elements. However, their widespread application remains hindered by challenges such as anisotropic crystal growth, composition control and lack

Travis Davies, Yiqi Huang, Alexi Gladstone, Yunxin Liu

Implicit policies parameterized by generative models, such as Diffusion Policy, have become the standard for policy learning and Vision-Language-Action (VLA) models in robotics. However, these approaches often suffer from high computational cost, exposure bias, and unstable inference dynamics, which lead to divergence under distribution shifts. Energy-Based

Jan Strich, Adeline Scharfenberg, Chris Biemann, Martin Semmann

We introduce EncouRAGe, a comprehensive Python framework designed to streamline the development and evaluation of Retrieval-Augmented Generation (RAG) systems using Large Language Models (LLMs) and Embedding Models. EncouRAGe comprises five modular and extensible components: Type Manifest, RAG Factory, Inference, Vector Store, and Metrics, facilitating flexi

Malik H. Altakrori, Nizar Habash, Abed Alhakim Freihat, Younes Samih

We present DialectalArabicMMLU, a new benchmark for evaluating the performance of large language models (LLMs) across Arabic dialects. While recently developed Arabic and multilingual benchmarks have advanced LLM evaluation for Modern Standard Arabic (MSA), dialectal varieties remain underrepresented despite their prevalence in everyday communication. Dialec

Naomi Muggleton, Timothy Monteath, Taha Yasseri

This study explores visitor behaviour at The British Museum using data science methods applied to novel sources, including audio guide usage logs and TripAdvisor reviews. Analysing 42,000 visitor journeys and over 50,000 reviews, we identify key drivers of satisfaction, segment visitors by behavioural patterns, examine tour engagement, model spatial navigati

A. V. Anufriev, V. N. Kovalenko

In this paper, we discuss the matching of the holographic equation of state with the equation of Hadron Resonance Gas for studying the nuclear matter properties within the framework of relativistic heavy-ion collisions. Machine learning methods are applied to the calibration of model's free parameters using the lattice QCD results for the physical values of

Kira van Treek, Javier F. Peña, Juan C. Vera, Luis F. Zuluaga

Many optimization algorithms$\unicode{x2013}$including gradient descent, proximal methods, and operator splitting techniques$\unicode{x2013}$can be formulated as fixed-point iterations (FPI) of continuous operators. When these operators are averaged, convergence to a fixed point is guaranteed when one exists, but the convergence is generally sublinear. Recen

Jamieson A L Howard, Benjamin Ambrose, Mahmoud A S Abdelhamid, Lewis Frame

DNA helicases undergo conformational changes; however, their structural dynamics are poorly understood. Here, we study single molecules of superfamily 1A DNA helicase Rep, which undergo conformational transitions during bacterial DNA replication, repair and recombination. We use time-correlated single-photon counting (TCSPC), fluorescence correlation spectro

King L. Ng, Luís H. Carnevale, Michał Klamka, Piotr Deuar

Droplet deformations caused by substrate vibrations are ubiquitous in nature and highly relevant for applications such as microreactors and single-cell sorting. The vibrations can induce droplet oscillations, a fundamental process that requires an in-depth understanding. Here, we report on extensive many-body dissipative particle dynamics simulations carried

Mohammad Zahangir Alam, Khandoker Ashik Uz Zaman, Mahdi H. Miraz

Retrieval-Augmented Generation (RAG) shows significant promise in knowledge-intensive tasks by improving domain specificity, enhancing temporal relevance, and reducing hallucinations. However, applying RAG to finance encounters critical challenges: restricted access to proprietary datasets, limited retrieval accuracy, regulatory constraints, and sensitive da

Pedro J. Colmenares

This research presents a framework for describing the average velocity magnitude of an accelerated, self-propelled Brownian particle diffusing in a thermal fluid and confined by a harmonic external potential. The system is immersed in a thermal bath of harmonic oscillators at a constant temperature, where the bath constituents also interact with the external

Maria Lizarazo Jimenez, Ana Gabriela Claros, Kieran Green, David Toro-Tobon

Large Language Models (LLMs) are increasingly demonstrating the potential to reach human-level performance in generating clinical summaries from patient-clinician conversations. However, these summaries often focus on patients' biology rather than their preferences, values, wishes, and concerns. To achieve patient-centered care, we propose a new standard for

Yue-Yang Fei, Zhenhuan Liu, Rui Zhang, Zhenyu Cai

Quantum networks, which integrate multiple quantum computers and the channels connecting them, are crucial for distributed quantum information processing but remain inherently susceptible to channel noise. Channel purification emerges as a promising technique for suppressing noise in quantum channels without complex encoding and decoding operations, making i

Khandoker Ashik Uz Zaman, Mohammad Zahangir Alam, Mohammed N. M. Ali, Mahdi H. Miraz

The protection of intellectual property has become critical due to the rapid growth of three-dimensional content in digital media. Unlike traditional images or videos, 3D point clouds present unique challenges for copyright enforcement, as they are especially vulnerable to a range of geometric and non-geometric attacks that can easily degrade or remove conve

Neha Srikanth, Victor Bursztyn, Puneet Mathur, Ani Nenkova

We introduce SQLSpace, a human-interpretable, generalizable, compact representation for text-to-SQL examples derived with minimal human intervention. We demonstrate the utility of these representations in evaluation with three use cases: (i) closely comparing and contrasting the composition of popular text-to-SQL benchmarks to identify unique dimensions of e

Massimiliano Puglisi

This thesis revolves around the Stolz' positive scalar curvature sequence: in particular adapted to the context of (G, F)-spaces, i.e. proper G-spaces with isotropy groups belonging to a family F of subgroups of G, and to that of manifolds with non-isolated singularities. In both cases, the sequence is studied for appropriate classes of metrics with positive

Jin Qin, Benedikt Schurr, Patrick Pertsch, Daniel Friedrich

Strong coupling between quantum emitters and optical cavities is essential for quantum information processing, high-purity single-photon sources, and nonlinear quantum devices. Achieving this regime at room temperature in a compact, deterministic on-chip platform-critical for integration with nanoelectronic circuitry and scalable device architectures-remains

A. V. Bomediano, R. J. Conanan, L. D. Santuyo, A. Coronel

Understanding the structural and cognitive underpinnings of musical compositions remains a key challenge in music theory and computational musicology. While traditional methods focus on harmony and rhythm, cognitive models such as the Implication-Realization (I-R) model and Temporal Gestalt theory offer insight into how listeners perceive and anticipate musi

Valentin Fonck, Mohammadali Razeghi, Jean Spièce, Phillip Dobson

Efficient thermal management is critical for cryogenic CMOS circuits, where local heating can compromise device performance and qubit coherence. Understanding heat flow at the nanoscale in these multilayer architectures requires localized, high-resolution thermal probing techniques capable of accessing buried structures. Here, we introduce a sideband thermal

Gabriel D. Patrón, Di Zhang, Lavinia M. P. Ghilardi, Evelin Blom

Energy storage promotes the integration of renewables by operating with charge and discharge policies that balance an intermittent power supply. A key challenge in this emerging sector is how to optimize the operation of storage assets given future price uncertainties and the need to recover the costs of project finance while ensuring an attractive return on

Yuxiang Chen, Yifan Liu, Xiaoming Xu, Pengle Zhang

Large Language Models (LLMs) training is prohibitively expensive, driving interest in low-precision fully-quantized training (FQT). While novel 4-bit formats like NVFP4 offer substantial efficiency gains, achieving near-lossless training at such low precision remains challenging. We introduce TetraJet-v2, an end-to-end 4-bit FQT method that leverages NVFP4 f

J. Poole, N. Dervilis, K. Worden, P. Gardner

Data for training structural health monitoring (SHM) systems are often expensive and/or impractical to obtain, particularly for labelled data. Population-based SHM (PBSHM) aims to address this limitation by leveraging data from multiple structures. However, data from different structures will follow distinct distributions, potentially leading to large genera

Christian Bick, Bob W. Rink, Babette A. J. de Wolff

Network interactions between dynamical units are often subject to time delay. We develop a phase reduction method for delay-coupled oscillator networks. The method is based on rewriting the delay-differential equation as an ordinary differential equation coupled with a transport equation, expanding in the coupling strength, and solving the resulting equation

Manzi Huang, Xiantao Wang, Zhuang Wang, Zhihao Xu

In this paper, we establish a quantitative correspondence between power quasi-symmetric mappings on complete metric spaces and rough quasi-isometric mappings on their hyperbolic fillings. In particular, we prove that the exponents in the power quasi-symmetric mappings coincide with the coefficients in the rough quasi-isometric mappings. This shows that the o

Théo Gnassounou, Yessin Moakher, Shifeng Xie, Vasilii Feofanov

Foundation models for time series are emerging as powerful general-purpose backbones, yet their potential for domain-specific biomedical signals such as electroencephalography (EEG) remains rather unexplored. In this work, we investigate the applicability a recently proposed time series classification foundation model, to a different EEG tasks such as motor

Nick Judd, Alexandre Vaz, Kevin Paeth, Layla Inés Davis

We introduce findings and methods to facilitate evidence-based discussion about how large language models (LLMs) should behave in response to user signals of risk of suicidal thoughts and behaviors (STB). People are already using LLMs as mental health resources, and several recent incidents implicate LLMs in mental health crises. Despite growing attention, f

Manuel Fernando Cortez, Oscar Jarrin

We consider a new nonlocal and nonlinear one-dimensional evolution model arising in the study of oceanic flows in equatorial regions, recently derived in [A. Constantin and L. Molinet, Global Existence and Finite-Time Blow-Up for a Nonlinear Nonlocal Evolution Equation, Commun. Math. Phys. 402 (2023), 3233-3252]. We investigate the spatial asymptotic behavio

Shinichi Nakagawa, Ayumi Mizuno, Coralie Williams, Santiago Ortega

Meta-analyses require an effect-size estimate and its corresponding sampling variance from primary studies. In some cases, estimators for the sampling variance of a given effect size statistic may not exist, necessitating the derivation of a new formula for sampling variance. Traditionally, sampling variance formulas are obtained via hand-derived Taylor expa

Zherui Yang, Zhehao Li, Kangbo Lyu, Yixuan Li

The conjugate gradient solver (CG) is a prevalent method for solving symmetric and positive definite linear systems Ax=b, where effective preconditioners are crucial for fast convergence. Traditional preconditioners rely on prescribed algorithms to offer rigorous theoretical guarantees, while limiting their ability to exploit optimization from data. Existing

Desta Haileselassie Hagos, Legand L. Burge, Anietie Andy, Anis Yazidi

Transformer-based architectures have advanced text summarization, yet their quadratic complexity limits scalability on long documents. This paper introduces BiSparse-AAS (Bilinear Sparse Attention with Adaptive Spans), a novel framework that combines sparse attention, adaptive spans, and bilinear attention to address these limitations. Sparse attention reduc

Yongjie Liu, Gangshan Jing, Long Wang

Graph rigidity theory is an important tool for examining the solvability of sensor network localization (SNL) problems, and ensuring global convergence of localization algorithms. Along this direction, diverse measurements such as signed angle (SA) and ratio of distance (RoD) have been considered. However, little is known about how the bipartition of nodes b

Valerio Assenza, Giorgia Testolina

In the setting of electromagnetic systems, we propose a new definition of electromagnetic Ricci curvature, naturally derived via the classical Jacobi-Maupertuis reparametrization from the recent works of Assenza [IMRN, 2024] and Assenza, Marshall Reber, Terek [Communications in Mathematical Physics, 2025]. On closed manifolds, we show that if the magnetic fo

Francisco Javier Escoto López

In this PhD thesis, a method for solving fast and accurately the monoenergetic drift-kinetic equation at low collisionality is presented. The algorithm is based on the analytical properties of the drift-kinetic equation when its dependence on the pitch-angle cosine is represented employing Legendre polynomials as basis functions. The Legendre representation

Mahi Aliyu Aminu, Chisom Chibuike, Fatimo Adebanjo, Omokolade Awosanya

Machine learning models often fail under distribution shifts, a problem exacerbated in low-resource settings where limited data restricts robust generalization. Domain generalization(DG) methods address this challenge by learning representations that remain invariant across domains, frequently leveraging causal principles. In this work, we study two causal D

Chunyin Chen, Sizhe Yan, Biao Wu

Entanglement entropy is a fundamental diagnostic for quantum chaos, typically exhibiting volume-law scaling in highly excited eigenstates of chaotic many-body systems. In this work, we present a striking counterexample: a Floquet-driven quantum many-body system with Rydberg-like blockade that, despite being fully chaotic as indicated by its Wigner-Dyson leve

A. A. Raduta, C. M. Raduta, R. Poenaru, Al. H. Raduta

A two interacting rotors Hamiltonian is alternatively treated semi-classically and by a Dyson boson expansion method. The linearized equations of motion lead to dispersion equation for the wobbling frequency. One defined a ground band with energies consisting in a rotational part and one half of the vibrational wobbling energy. Adding to each state energy th

Julien Philip, Li Ma, Pascal Clausen, Wenqi Xian

We present a unique system for large-scale, multi-performer, high resolution 4D volumetric capture providing realistic free-viewpoint video up to and including 4K resolution facial closeups. To achieve this, we employ a novel volumetric capture, reconstruction and rendering pipeline based on Dynamic Gaussian Splatting and Diffusion-based Detail Enhancement.

Rinaldo M. Colombo, Mauro Garavello, Claudia Nocita

We present a macroscopic traffic flow model where standard vehicles coexist with vehicles informed on the traffic distribution. The resulting mixed nonlocal-local integro-differential PDEs is proved to generate a locally Lipschitz continuous semigroup whose orbits are uniquely characterized as solutions to the system, according to a natural definition of sol

Elena Mulero Ayllón, Linlin Shen, Pierangelo Veltri, Fabrizia Gelardi

Accurate lung tumor segmentation is vital for improving diagnosis and treatment planning, and effectively combining anatomical and functional information from PET and CT remains a major challenge. In this study, we propose vMambaX, a lightweight multimodal framework integrating PET and CT scan images through a Context-Gated Cross-Modal Perception Module (CGM

Ke He, Thang X. Vu, Le He, Lisheng Fan

The rise of ultra-dense LEO constellations creates a complex and asynchronous network environment, driven by their massive scale, dynamic topologies, and significant delays. This unique complexity demands an adaptive packet routing algorithm that is asynchronous, risk-aware, and capable of balancing diverse and often conflicting QoS objectives in a decentral

T. Gayoux, P. -S. Corasaniti, T. R. G. Richardson, S. T. Kay

The large catalogues of galaxy clusters expected from the Euclid survey will enable cosmological analyses of cluster number counts that require accurate cosmological model predictions. One possibility is to use parametric fits calibrated against $N$-body simulations, that capture the cosmological parameter dependence of the halo mass function. Several studie

Junkang Liu, Yuxuan Tian, Fanhua Shang, Yuanyuan Liu

To prevent inference attacks in Federated Learning (FL) and reduce the leakage of sensitive information, Client-level Differentially Private Federated Learning (CL-DPFL) is widely used. However, current CL-DPFL methods usually result in sharper loss landscapes, which leads to a decrease in model generalization after differential privacy protection. By using

Anubhab Ghosh, Yonina C. Eldar, Saikat Chatterjee

We consider the problem of designing a data-driven nonlinear state estimation (DANSE) method that uses (noisy) nonlinear measurements of a process whose underlying state transition model (STM) is unknown. Such a process is referred to as a model-free process. A recurrent neural network (RNN) provides parameters of a Gaussian prior that characterize the state

Daniel Lozano-Martín, Dirk Tuma, César R. Chamorro

This study evaluates the performance of three reference equations of state (EoS), AGA8-DC92, GERG-2008, and SGERG-88, in predicting the density of regasified liquefied natural gas (RLNG) mixtures. A synthetic nine-component RLNG mixture was gravimetrically prepared. High-precision density measurements were obtained using a single-sinker magnetic suspension d

María Mallo, Marcelino Agúndez, Carlos Cabezas, Octavio Roncero

Cyclopentadiene (c-C5H6) is considered a key molecule in the formation of polycyclic aromatic hydrocarbons (PAHs) in the interstellar medium (ISM). The synthesis of PAHs from simpler precursors is known as the "bottom-up" theory, which, so far, has been dominated by reactions between organic radicals. However, this mechanism struggles to account for the orig

Dennis Trautwein, Cornelius Ihle, Moritz Schubotz, Bela Gipp

The promise of decentralized peer-to-peer (P2P) systems is fundamentally gated by the challenge of Network Address Translation (NAT) traversal, with existing solutions often reintroducing the very centralization they seek to avoid. This paper presents the first large-scale, longitudinal measurement study of a fully decentralized NAT traversal protocol, Direc

Michaël De San Féliciano, Christophe Adessi, Julien El Hajj, Nicolas Horny

With the increasing miniaturization of electronic components and the need to optimize thermal management, it has become essential to understand heat transport at metal/semiconductor interfaces. While it has been recognized decades ago that an electron phonon channel may take place at metal-semiconductor interfaces, its existence is still controversial. Here,

Binh Thuan Tran, Nicolas Schreuder

We study the problem of nonparametric two-sample testing using the sliced Wasserstein (SW) distance. While prior theoretical and empirical work indicates that the SW distance offers a promising balance between strong statistical guarantees and computational efficiency, its theoretical foundations for hypothesis testing remain limited. We address this gap by

Liam Kearns

The integration of AI tools into medical applications has aimed to improve the efficiency of diagnosis. The emergence of large language models (LLMs), such as ChatGPT and Claude, has expanded this integration even further despite a concern for their environmental impact. Because of LLM versatility and ease of use through APIs, these larger models are often u

Ian Koot, C. J. F. van de Ven

The aim of this paper is two-fold. First, we prove the existence of Lieb-Robinson bounds for classical particle systems describing harmonic oscillators interacting with arbitrarily many neighbors, both on lattices and on more general structures. Second, we prove the existence of a global dynamical system on the commutative resolvent algebra, a C*-algebra of

Elena Pinilla-Cienfuegos, Lucas Mascaró-Burguera, Ramón Torres-Cavanillas, J. Ignacio Echavarría

The ability to control and understand the phase transitions of individual nanoscale building blocks is key to advancing the next generation of low-power reconfigurable nanophotonic devices. To address this critical challenge, molecular nanoparticles (NPs) exhibiting a spin crossover (SCO) phenomenon are trapped by coupling a quadrupole Paul trap with a multi

Antonio Desiderio, Alessia Galdeman, Franziska Bauerlein, Sune Lehmann

Food waste represents a major challenge to global climate resilience, accounting for almost 10% of annual greenhouse gas emissions. The retail sector is a critical player, mediating product flows between producers and consumers, where supply chain inefficiencies can shape which items are put on sale. Yet how these dynamics vary across geographic contexts rem

Eunku Park, Antoine Vigneron

We show how to construct in linear time coresets of constant size for farthest point problems in fixed-dimensional hyperbolic space. Our coresets provide both an arbitrarily small relative error and additive error $\varepsilon$. More precisely, we are given a set $P$ of $n$ points in the hyperbolic space $\mathbb{H}^D$, where $D=O(1)$, and an error tolerance

Gibson De Marchi Poltronieri, Farney C. Moreira, João Luiz F. Azevedo

Numerical simulations of reactive hypersonic flow under thermodynamic and chemical non-equilibrium conditions are presented for the Mars Pathfinder capsule. An 8-species chemical model is employed to simulate Mars' atmosphere. Park's two-temperature model is used to account for the thermal non-equilibrium phenomena. The present work analyzes the impact of di

Jonathan Martinez, Teresa Radice, Francesco Sorrentino

This work introduces a systematic algorithm for generating directed networks with prescribed symmetries by constructing expansions from a given quotient network. The method enables researchers to synthesize realistic network models with controllable symmetry structure, facilitating studies of symmetry-driven dynamics such as cluster synchronization in biolog

Anssi Manninen, Janek Gröhl, Felix Lucka, Andreas Hauptmann

Photoacoustic tomography (PAT) is a medical imaging modality that can provide high-resolution tissue images based on the optical absorption. Classical reconstruction methods for quantifying the absorption coefficients rely on sufficient prior information to overcome noisy and imperfect measurements. As these methods utilize computationally expensive forward

Qing Guo, Xinhang Li, Junyu Chen, Zheng Guo

Leveraging large language models (LLMs) in traffic signal control (TSC) improves optimization efficiency and interpretability compared to traditional reinforcement learning (RL) methods. However, existing LLM-based approaches are limited by fixed time signal durations and are prone to hallucination errors, while RL methods lack robustness in signal timing de

Junkang Liu, Fanhua Shang, Hongying Liu, Yuxuan Tian

AdamW has become one of the most effective optimizers for training large-scale models. We have also observed its effectiveness in the context of federated learning (FL). However, directly applying AdamW in federated learning settings poses significant challenges: (1) due to data heterogeneity, AdamW often yields high variance in the second-moment estimate $\

Ben Fiedler, Samuel Gruetter, Timothy Roscoe

The ever increasing complexity of hardware platforms poses a challenge to systems programmers. Correctly programming a multitude of components, providing functionality and security, is difficult: semantics of individual units are described in prose, underspecified, and prone to inaccuracies. Rigorous statements about platform security are often impossible. W

Uzay Macar, Paul C. Bogdan, Senthooran Rajamanoharan, Neel Nanda

Most work interpreting reasoning models studies only a single chain-of-thought (CoT), yet these models define distributions over many possible CoTs. We argue that studying a single sample is inadequate for understanding causal influence and the underlying computation. Though fully specifying this distribution is intractable, we can measure a partial CoT's im

Lei Geng, Sujay Ray, Philipp Werner

Multiorbital Hubbard models with Hund coupling and crystal-field splitting exhibit an instability toward spin-triplet excitonic order in the parameter regime characterized by strong local spin fluctuations. Upon chemical doping, two distinct types of excitonic ferromagnetism have been reported. Using steady-state nonequilibrium dynamical mean-field theory, w

Wei Xu, Cheng Wang, Dingkang Liang, Zongchuang Zhao

Underwater exploration offers critical insights into our planet and attracts increasing attention for its broader applications in resource exploration, national security, etc. We study the underwater scene understanding methods, which aim to achieve automated underwater exploration. The underwater scene understanding task demands multi-task perceptions from

Bernardo Williams, Victor M. Yeom-Song, Marcelo Hartmann, Arto Klami

We propose a method for learning and sampling from probability distributions supported on the simplex. Our approach maps the open simplex to Euclidean space via smooth bijections, leveraging the Aitchison geometry to define the mappings, and supports modeling categorical data by a Dirichlet interpolation that dequantizes discrete observations into continuous

María del Carmen Romero, Mariana del Fresno, Alejandro Clausse

This paper addresses the challenge of identifying a minimal subset of discrete, independent variables that best predicts a binary class. We propose an efficient iterative method that sequentially selects variables based on which one provides the most statistically significant reduction in conditional entropy, using confidence bounds to account for finite-sam

Swarang Joshi

This study investigates how well computational embeddings align with human semantic judgments in the processing of English compound words. We compare static word vectors (GloVe) and contextualized embeddings (BERT) against human ratings of lexeme meaning dominance (LMD) and semantic transparency (ST) drawn from a psycholinguistic dataset. Using measures of a

Lorenz J. J. Sauerzopf, Fabian Becker, Kai Müller

Photonic integrated circuits are heavily researched devices for telecommunication, biosensing, and quantum technologies. Wafer-scale fabrication and testing are crucial for reducing costs and enabling large-scale deployment. Grating couplers allow non-invasive measurements before packaging, but classical designs rely on long tapers and narrow bandwidths. In

Hyemin Boo, Eunsang Lee, Jiyoung Lee

Deepfakes generated by advanced generative models have rapidly posed serious threats, yet existing audiovisual deepfake detection approaches struggle to generalize to unseen manipulation methods. To address this, we propose a novel reference-aware audiovisual deepfake detection method, called Referee to capture fine-grained identity discrepancies. Unlike exi

Angelica Liguori, Ettore Ritacco, Pietro Sabatino, Annalisa Socievole

Graphs are central to modeling complex systems in domains such as social networks, molecular chemistry, and neuroscience. While Graph Neural Networks, particularly Graph Convolutional Networks, have become standard tools for graph learning, they remain constrained by reliance on fixed structures and susceptibility to over-smoothing. We propose the Spectral P

Duaa Abdullah, Jasem Hamoud

In this paper we explore fundamental concepts in computational complexity theory and the boundaries of algorithmic decidability. We examine the relationship between complexity classes \textbf{P} and \textbf{NP}, where $L \in \textbf{P}$ implies the existence of a deterministic Turing machine solving $L$ in polynomial time $O(n^k)$. Central to our investigati

Carles Roch I Carceller, Armin Tavakoli

A promising platform for semi-device-independent quantum information is prepare-and-measure experiments restricted only by a bound on the energy of the communication. Here, we investigate the role of shared entanglement in such scenarios. For classical communication, we derive a general correlation criterion for nonlocal resources and use it to show that ent

Xylo Molenda, S. Zhong, B. Viswanathan, Xingli Li

Extending classical synchronization to the quantum domain is of great interest both from the fundamental physics point of view and with a view toward quantum technology applications. This work characterizes phase synchronization of an effective spin-1 system, which is realized by coupling three quantum states with infinite lifetime to auxiliary excited state

Einius Pultinevicius, Aaron Götzelmann, Fabian Thielemann, Christian Hölzl

Stability achieved by large angular momentum is ubiquitous in nature, with examples ranging from classical mechanics, over optics and chemistry, to nuclear physics. In atoms, angular momentum can protect excited electronic orbitals from decay due to selection rules. This manifests spectacularly in highly excited Rydberg states. Low angular momentum Rydberg s

Veselin B. Kostov, Brian P. Powell, Saul A. Rappaport, Tamas Borkovits

We present the discovery of 52 eclipsing quadruple star candidates detected in TESS Full Frame Image eleanor data by machine learning and citizen scientists. The uniformly-vetted and -validated targets exhibit two sets of eclipses following two distinct periods, representing quadruple systems with a 2+2 hierarchical configuration. Detailed photocenter measur

Chenyang Shao, Sijian Ren, Fengli Xu, Yong Li

In recent years, large language models (LLMs) have witnessed remarkable advancements, with the test-time scaling law consistently enhancing the reasoning capabilities. Through systematic evaluation and exploration of a diverse spectrum of intermediate thoughts, LLMs demonstrate the potential to generate deliberate reasoning steps, thereby substantially enhan

Ameemah Humayun, Bushra Zubair, Maryam Mustafa

Low-resource countries represent over 90% of maternal deaths, with Pakistan among the top four countries contributing nearly half in 2023. Since these deaths are mostly preventable, large language models (LLMs) can help address this crisis by automating health communication and risk assessment. However, sexual and reproductive health (SRH) communication in c

Yu Wang, Chris W. Ormel, Hao-Chang Jiang, Sebastiaan Krijt

V883 Ori is an FU-Orionis-type outburst system characterized by a shoulder at 50-70 au in its ALMA band 6 and 7 intensity profiles. Previously, this feature was attributed to dust pile-up from pebble disintegration at the water snowline. However, recent multi-wavelength observations show continuity in the spectral index across the expected snowline region, d

Daniele Cirulli, Antonio Desiderio, Giulio Cimini, Fabio Saracco

Political debate nowadays takes place mainly on online social media, with election periods amplifying ideological engagement. Reddit is generally considered more resistant to polarization and echo chamber effects than platforms like Twitter or Facebook. Here, we challenge this assumption through a case study across the 2016 US presidential election. We use s

Felix George, Harshit Kumar, Divya Pathak, Kaustabha Ray

Agentic applications powered by Large Language Models exhibit non-deterministic behaviors that can form hidden execution cycles, silently consuming resources without triggering explicit errors. Traditional observability platforms fail to detect these costly inefficiencies. We present an unsupervised cycle detection framework that combines structural and sema

Aslan Ahadi, Elham Sarvari, Jan Frenzel, Gunther Eggeler

The dependence of transformation pattern in superelastic NiTi tubes on tube outer diameter D and wall-thickness t is investigated through quasi-static uniaxial tension and large-rotation bending experiments. The evolution of outer-surface strain fields is synchronized with global stress-strain and moment-curvature responses using a multi-magnification, high-

Tamas Gombor, Adolfo Holguin

We consider $\mathfrak{so}_4$ invariant matrix product states (MPS) in the $\mathfrak{so}_6$ symmetric integrable spin chain and prove their integrability. These MPS appear as fuzzy three-sphere solutions of matrix models with Yang-Mills-type interactions, and in particular they correspond to scalar defect sectors of $N=4$ SYM. We find that the algebra forme