March 2025 arXiv papers — page 38

Tommaso Tonolo, Maria Chiara Angelini, Sandro Azaele, Amos Maritan

We study the equilibrium phases of a generalized Lotka-Volterra model characterized by a species interaction matrix which is random, sparse and symmetric. Dynamical fluctuations are modeled by a demographic noise with amplitude proportional to the effective temperature T. The equilibrium distribution of species abundances is obtained by means of the cavity m

Mateo Pascual

We generalise Langlois' Hamiltonian treatment of gauge-invariant linear cosmological perturbations to a cosmological setting with multiple scalar fields minimally coupled to gravity. We review the Hamilton-Jacobi-like technique for a Hamiltonian system with first-class constraints. With this technique, elucidating the gauge-invariant quantities of the system

Imogen G. Cresswell, Adrian E. Fraser, Evan B. Bauer, Evan H. Anders

Polluted white dwarfs (WDs) with small surface convection zones deposit significant concentrations of heavy elements to the underlying radiative interior, presumably driving thermohaline convection. Current models of polluted WDs frequently fail to account for this effect, although its inclusion can increase the inferred accretion rate by orders of magnitude

Usama Zafar, André M. H. Teixeira, Salman Toor

Federated learning (FL) enables collaborative model training across distributed clients without sharing raw data, but its robustness is threatened by Byzantine behaviors such as data and model poisoning. Existing defenses face fundamental limitations: robust aggregation rules incur error lower bounds that grow with client heterogeneity, while detection-based

Nairen Cao, Vincent Cohen-Addad, Shi Li, Euiwoong Lee

Correlation Clustering is a fundamental and widely-studied problem in unsupervised learning and data mining. The input is a graph and the goal is to construct a clustering minimizing the number of inter-cluster edges plus the number of missing intra-cluster edges. CCL+24 introduced the cluster LP for Correlation Clustering, which they argued captures the pro

Max Griswold, Beth Ann Griffin, Max Rubinstein, Mincen Liu

Using state-level opioid overdose mortality data from 1999-2016, we simulated four time-varying treatment scenarios, which correspond to real-world policy dynamics (ramp up, ramp down, temporary and inconsistent). We then evaluated seven commonly used policy evaluation methods: two-way fixed effects event study, debiased autoregressive model, augmented synth

Mustafa A. Amin, M. Sten Delos, Mehrdad Mirbabayi

We investigate the evolution of density perturbations in dark matter, including the new combined effects of finite number density and non-zero velocity dispersion. Using a truncated BBGKY hierarchy, we derive analytical expressions for the dark matter power spectrum during radiation and matter domination. A component of warm white noise emerges in our analys

Amaya Gallagher-Syed, Henry Senior, Omnia Alwazzan, Elena Pontarini

The development of biologically interpretable and explainable models remains a key challenge in computational pathology, particularly for multistain immunohistochemistry (IHC) analysis. We present BioX-CPath, an explainable graph neural network architecture for whole slide image (WSI) classification that leverages both spatial and semantic features across mu

Laura Lewis, Dar Gilboa, Jarrod R. McClean

Without large quantum computers to empirically evaluate performance, theoretical frameworks such as the quantum statistical query (QSQ) are a primary tool to study quantum algorithms for learning classical functions and search for quantum advantage in machine learning tasks. However, we only understand quantum advantage in this model at two extremes: either

Alexander Beckett, Marc Rafelski, Claudia Scarlata, Wanjia Hu

We present the survey design and initial results from the Parallel Ionizing Emissivity (PIE) survey. PIE is a large HST survey designed to detect Lyman continuum (LyC) emitting galaxies at 3.1$<$ z $<$3.5 and stack their images in order to measure average LyC escape fractions as a function of galaxy properties. PIE has imaged 37 independent fields in three f

Léonard Lehoucq, Irina Dvorkin, Luciano Rezzolla

The stochastic gravitational-wave background (SGWB) generated by the inspiral and merger of binary neutron stars is traditionally modelled assuming that the inspiral is promptly followed by the collapse of the merger remnant to a rotating black hole. While this is reasonable for the most massive binaries, it is not what is expected in general, as the remnant

F. Rizzuti, G. Cescutti, P. Molaro, L. Roberti

Recent campaigns of observations have provided new measurements of the carbon isotopes in the most metal-poor stars of the Galaxy. These stars are so metal-poor that they could only have been enriched by one or few generations of massive progenitors. However, explaining the primary production of $^{13}$C and the low $^{12}$C/$^{13}$C ratio measured in these

Leigh C. Smith, Saad Ahmed, Francesca De Angeli, P. W. Burgess

We present the Cambridge Exoplanet Transit Recovery Algorithm (CETRA), a fast and sensitive transit detection algorithm, optimised for GPUs. CETRA separates the task into a search for transit signals across linear time space, followed by a phase-folding of the former to enable a periodic signal search, using a physically motivated transit model to improve de

S. B. Kožić, G. Torre, K. Delić, F. Franchini

The study of entanglement and magic properties in topologically frustrated systems suggests that, in the thermodynamic limit, these quantities decompose into two distinct contributions. One is determined by the specific nature of the model and its Hamiltonian, and another arises from topological frustration itself, resulting in being independent of the Hamil

Daniele De Bernardis, Hugo Levy-Falk, Elena Fanella, Rocco Duquennoy

We theoretically investigate a single fluorescent molecule as a hybrid quantum optical device, in which multiple external laser sources exert control of the vibronic states. In the high-saturation regime, a coherent interaction is established between the vibrational and electronic degrees of freedom, and molecules can simulate several cavity QED models, wher

Silin Gao, Sheryl Mathew, Li Mi, Sepideh Mamooler

Visual narrative generation transforms textual narratives into sequences of images illustrating the content of the text. However, generating visual narratives that are faithful to the input text and self-consistent across generated images remains an open challenge, due to the lack of knowledge constraints used for planning the stories. In this work, we propo

Reza Haghshenas, Eli Chertkov, Michael Mills, Wilhelm Kadow

Digital quantum matter -- realized when discrete quantum gates approximate continuous time evolution -- is susceptible to heating into chaotic, structureless states. If digitization errors are adequately suppressed, a long-lived transient regime of approximately energy-conserving dynamics can be observed on gate-based quantum computers. Conservation of energ

Rimpei Chiba, Neige Frankel, Chris Hamilton

Gaia recently revealed a two-armed spiral pattern in the vertical phase-space distribution of the inner Galactic disk (guiding radius $R_\textrm{g} \sim 6.2$ kpc), indicating that some non-adiabatic perturbation symmetric about the mid-plane is driving the inner disk out of equilibrium. The non-axisymmetric structures in the disk (e.g., the bar or spiral arm

Michele Martone, Julia Lawall

Currently, the most energy-efficient hardware platforms for floating point-intensive calculations (also known as High Performance Computing, or HPC) are graphical processing units (GPUs). However, porting existing scientific codes to GPUs can be far from trivial. This article summarizes our recent advances in enabling machine-assisted, HPC-oriented refactori

Junze Deng, Yi Jiang, Tiago F. T. Cerqueira, Haoyu Hu

We present a comprehensive investigation of the flat-band kagome superconductor LaRu$_3$Si$_2$, which has recently been reported to host charge density wave (CDW) order above room temperature ($T_{CDW} \simeq 400$ K). The stable crystal structure above the CDW transition is identified via soft phonon condensation and confirmed to be harmonically stable throu

Pietro Benetti Genolini, Sameer Murthy

We study the Kontsevich-Segal-Witten criterion for allowable complex metrics, in the context of the gravitational path integral corresponding to the supersymmetric index. In various theories of supergravity in asymptotically flat and asymptotically AdS space, the exponential growth of states of the corresponding microscopic index in string theory is known to

Guillermo Arias-Tamargo, Chris Hull, Maxwell L. Velásquez Cotini Hutt

Global symmetries can be generalised to transformations generated by topological operators, including cases in which the topological operator does not have an inverse. A family of such topological operators are intimately related to dualities via the procedure of half-space gauging. In this work we discuss the construction of non-invertible defects based on

Camila de Sá-Freitas, Dimitri A. Gadotti, Francesca Fragkoudi, Paula Coelho

Once galaxies settle their discs and become self-gravitating, stellar bars can form, driving the subsequent evolution of their host galaxy. Determining the ages of bars can therefore shed light on the epoch of the onset of secular evolution. In this work, we apply the first broadly applicable methodology to derive bar ages to a sample of 20 nearby galaxies.

Shu-Heng Shao, Jonathan Sorce, Manu Srivastava

The algebraic approach to quantum field theory focuses on the properties of local algebras, whereas the study of (possibly non-invertible) global symmetries emphasizes global aspects of the theory and spacetime. We study connections between these two perspectives by examining how either of two core algebraic properties -- "additivity" or "Haag duality" -- is

F. M. Vincentelli, T. Shahbaz, P. Casella, V. S. Dhillon

We report on the detection of optical/near-infrared (O-IR) quasi-periodic oscillations (QPOs) from the black hole X-ray transient Swift J1727.8-1613. We obtained three X-ray and O-IR high-time-resolution observations of the source during its intermediate state (2023 September 9, 15 and 17) using NICER, HAWK-I@VLT, HIPERCAM@GTC and ULTRACAM@NTT. We clearly de

Miguel Antonio Sulangi, Willem Farmilo, Andreas Kreisel, Mainak Pal

Several recent experiments have challenged the premise that cuprate high-temperature superconductors approach conventional Landau-BCS behavior in the high-doping limit. We argue, based on an analysis of their superconducting spectra, that anomalous properties seen in the most-studied overdoped cuprates require a pairing interaction that is strongly inhomogen

Michalis Kourniotis, Lydia S. Cidale, Michaela Kraus, Matias A. Ruiz Diaz

Blue supergiants (BSGs) mediate between the main sequence and the late stages of massive stars, which makes them valuable for assessing the physics that drives the stars across the diverse evolutionary channels. By exploring correlations between the parameters of BSGs and their variability properties, we aim to improve the constraints on the models of the ev

Pavel A. Maksimov, Shengtao Jiang, L. P. Regnault, A. L. Chernyshev

The inelastic neutron scattering results and their analysis unequivocally point to a dominant Kitaev interaction in the honeycomb-lattice cobaltate BaCo$_2$(AsO$_4$)$_2$. Our anisotropic-exchange model closely describes $all$ available neutron scattering data in the material's field-polarized phase. The density-matrix renormalization group results for our mo

Silvio Fortuné, Rhea-Silvia Remus, Lucas C. Kimmig, Andreas Burkert

Our picture of galaxy evolution currently assumes that galaxies spend their life on the star formation main sequence (SFMS) until they are eventually quenched. However, recent observations show indications that the full picture might be more complicated. We reveal typical in-situ star formation histories and their relations to large-scale environment as well

Lucas C. Kimmig, Sarah Brough, Klaus Dolag, Rhea-Silvia Remus

As the most massive nodes of the cosmic web, galaxy clusters represent the best probes of structure formation. Over time, they grow by accreting and disrupting satellite galaxies, adding those stars to the brightest cluster galaxy (BCG) and the intra-cluster light (ICL). However, the formation pathways of different galaxy clusters can vary significantly. To

Guillermo Torres, Ralph Neuhäuser, Sebastian A. Hüttel, Valeri V. Hambaryan

Runaway stars are characterized by higher space velocities than typical field stars. They are presumed to have been ejected from their birth places by one or more energetic mechanisms, including supernova explosions. Accurate radial velocities are essential for investigating their origin, by tracing back their Galactic orbits to look for close encounters in

Jackson Yant, Miles Blencowe

We develop a quantum field-theoretic model of gravitationally induced entanglement (GIE) between two massive objects in spatial superposition. The masses are described as excitations of a scalar field in an external harmonic potential, allowing for a well-defined notion of relativistic coherent states. Using linearized quantum gravity in the static limit, we

Sunhaeng Hur, Vishnu Jejjala, Michael J. Kavic, Djordje Minic

In this paper we discuss possible consequences of a manifestly non-commutative and $T$-duality covariant formulation of string theory on dark energy, when the correspondence between short distance (UV) and long distance (IR) physics is taken into account. We demonstrate that the dark energy is dynamical, \textit{i.e.}, time-dependent, and we compute the allo

Sondos Mahmoud Bsharat, Mukul Ranjan, Aidar Myrzakhan, Jiacheng Liu

Rapid advancements in large language models (LLMs) have increased interest in deploying them on mobile devices for on-device AI applications. Mobile users interact differently with LLMs compared to desktop users, creating unique expectations and data biases. Current benchmark datasets primarily target at server and desktop environments, and there is a notabl

Alexander Swerdlow, Mihir Prabhudesai, Siddharth Gandhi, Deepak Pathak

Multimodal generative models that can understand and generate across multiple modalities are dominated by autoregressive (AR) approaches, which process tokens sequentially from left to right, or top to bottom. These models jointly handle images, text, video, and audio for various tasks such as image captioning, question answering, and image generation. In th

Tianqi Liu, Zihao Huang, Zhaoxi Chen, Guangcong Wang

We present Free4D, a novel tuning-free framework for 4D scene generation from a single image. Existing methods either focus on object-level generation, making scene-level generation infeasible, or rely on large-scale multi-view video datasets for expensive training, with limited generalization ability due to the scarcity of 4D scene data. In contrast, our ke

Jinwei Li, Huan-ang Gao, Wenyi Li, Haohan Chi

With the rapid advancements in diffusion models and 3D generation techniques, dynamic 3D content generation has become a crucial research area. However, achieving high-fidelity 4D (dynamic 3D) generation with strong spatial-temporal consistency remains a challenging task. Inspired by recent findings that pretrained diffusion features capture rich corresponde

Zichen Liu, Changyu Chen, Wenjun Li, Penghui Qi

DeepSeek-R1-Zero has shown that reinforcement learning (RL) at scale can directly enhance the reasoning capabilities of LLMs without supervised fine-tuning. In this work, we critically examine R1-Zero-like training by analyzing its two core components: base models and RL. We investigate a wide range of base models, including DeepSeek-V3-Base, to understand h

Yan-Bo Lin, Kevin Lin, Zhengyuan Yang, Linjie Li

In this paper, we introduce zero-shot audio-video editing, a novel task that requires transforming original audio-visual content to align with a specified textual prompt without additional model training. To evaluate this task, we curate a benchmark dataset, AvED-Bench, designed explicitly for zero-shot audio-video editing. AvED-Bench includes 110 videos, ea

Yulu Pan, Ce Zhang, Gedas Bertasius

We present BASKET, a large-scale basketball video dataset for fine-grained skill estimation. BASKET contains 4,477 hours of video capturing 32,232 basketball players from all over the world. Compared to prior skill estimation datasets, our dataset includes a massive number of skilled participants with unprecedented diversity in terms of gender, age, skill le

Lu Cheng, Qixiu Zhang, Beibei Xu, Zhiwei Huang

The transition to intelligent, low-carbon power systems necessitates advanced optimization strategies for managing renewable energy integration, energy storage, and carbon emissions. Generative Large Models (GLMs) provide a data-driven approach to enhancing forecasting, scheduling, and market operations by processing multi-source data and capturing complex s

Zhaorun Chen, Mintong Kang, Bo Li

Autonomous agents powered by foundation models have seen widespread adoption across various real-world applications. However, they remain highly vulnerable to malicious instructions and attacks, which can result in severe consequences such as privacy breaches and financial losses. More critically, existing guardrails for LLMs are not applicable due to the co

M. Siddikov

In this preprint we analyze the inclusive hadroproduction of heavy charmonia-bottomonia pairs in the Color Glass Condensate framework in the dilute-dense approximation. The production mechanisms can be classified by number of gluons emitted from projectile as Single- and Double Parton Scattering. We analyzed separately both types of contributions and evaluat

Michelle Guo, Matt Jen-Yuan Chiang, Igor Santesteban, Nikolaos Sarafianos

We introduce a novel approach to reconstruct simulation-ready garments with intricate appearance. Despite recent advancements, existing methods often struggle to balance the need for accurate garment reconstruction with the ability to generalize to new poses and body shapes or require large amounts of data to achieve this. In contrast, our method only requir

Lutz Mattner

For the usual normal approximations to binomial, hypergeometric, or Poisson interval probabilities, we collect some simple but then reasonably sharp error bounds. For the Clopper-Pearson~(1934) binomial confidence bounds, we present, following Michael Short's~(2023) approach, bounds similar to, but necessarily more complicated than, Lagrange's (1776) success

Hong Su, Baoyu Xu, Ju Chen, Chang Liu

The chiral gravitational wave background (GWB) can be produced by axion-like fields in the early universe. We perform parameter estimation for two types of chiral GWB with the LISA-Taiji network: axion-dark photon coupling and axion-Nieh-Yan coupling. We estimate the spectral parameters of these two mechanisms induced by axion and determine the normalized mo

M. Pálfi, G. Dálya, P. Raffai

Stellar mass can enhance the ranking of potential hosts for compact binary coalescences identified by ground-based gravitational-wave detectors within large localisation areas containing even thousands of galaxies. Despite its benefits, accurate stellar mass estimation is often time-consuming and computationally intensive. In this study, we implement four st

Shijie Zhou, Hui Ren, Yijia Weng, Shuwang Zhang

Recent advancements in 2D and multimodal models have achieved remarkable success by leveraging large-scale training on extensive datasets. However, extending these achievements to enable free-form interactions and high-level semantic operations with complex 3D/4D scenes remains challenging. This difficulty stems from the limited availability of large-scale,

Kerr Maxwell

We develop a geometric description of structured Gaussian beams, a form a structured light, by applying geometric quantisation and symplectic reduction to the 2D harmonic oscillator. Our results show that the geometric quantisation of the oscillator's reduced phase space coincides with the modal Poincar\'e sphere in optics. We explicitly consider the case of

Vipin Sudevan, Pisin Chen

We introduce a comprehensive, custom-developed neural network, the PUREPath-B, that yields a posterior predictive distribution of Cosmic Microwave Background (CMB) B-mode signal conditioned on the foreground contaminated CMB data and informed by the training dataset. Our network employs nested probabilistic multi-modal U-Net framework, enhanced with probabil

Orit Sela, Mary Schaps, Uzi Vishne

We consider quotients of the Bruhat-Tits building associated to the projective linear groups of dimension $d>2$ over the function field $\mathbb F_q(t)$ by a non-uniform lattice $\Gamma$ which is a congruence subgroup in the non-uniform lattice $ PGL_{d}(R)$, where $R=\mathbb F_q[\frac{1}{t}]$. We determine a fundamental domain and demonstrate that the quoti

Ilia Shilov, Ezzat Elokda, Sophie Hall, Heinrich H. Nax

Many multi-agent socio-technical systems rely on aggregating heterogeneous agents' costs into a social cost function (SCF) to coordinate resource allocation in domains like energy grids, water allocation, or traffic management. The choice of SCF often entails implicit assumptions and may lead to undesirable outcomes if not rigorously justified. In this paper

Masoumeh Sharafi, Emma Ollivier, Muhammad Osama Zeeshan, Soufiane Belharbi

Facial Expression Recognition (FER) from videos is a crucial task in various application areas, such as human-computer interaction and health diagnosis and monitoring (e.g., assessing pain and depression). Beyond the challenges of recognizing subtle emotional or health states, the effectiveness of deep FER models is often hindered by the considerable inter-s

E. Kammoun, I. E. Papadakis, M. Dovčiak, E. Lusso

It is established that the ultraviolet (UV) and X-ray emissions in active galactic nuclei (AGN) are tightly correlated. This correlation is observed both in low- and high-redshift sources. In particular, observations of large samples of quasars revealed the presence of a non-linear correlation between UV and X-rays. The physical origin of this correlation is

Ben Deaner, Chen-Wei Hsiang, Andrei Zeleneev

The presence of unobserved confounders is one of the main challenges in identifying treatment effects. In this paper, we propose a new approach to causal inference using panel data with large large $N$ and $T$. Our approach imputes the untreated potential outcomes for treated units using the outcomes for untreated individuals with similar values of the laten

Haotian Yang, Zhuoran Wang, Benson Chou, Sophie Xu

Federated Learning (FL) enables distributed ML model training on private user data at the global scale. Despite the potential of FL demonstrated in many domains, an in-depth view of its impact on model accuracy remains unclear. In this paper, we investigate, systematically, how this learning paradigm can affect the accuracy of state-of-the-art ML models for

Clara Fannjiang, Ji Won Park

Algorithms for machine learning-guided design, or design algorithms, use machine learning-based predictions to propose novel objects with desired property values. Given a new design task -- for example, to design novel proteins with high binding affinity to a therapeutic target -- one must choose a design algorithm and specify any hyperparameters and predict

Lorenzo Gavassino, Henry Hirvonen, Jean-François Paquet, Mayank Singh

The mean transverse momentum $\langle p_T \rangle$ of hadrons has been observed experimentally and in numerical simulations to have a power-law dependence on the hadronic multiplicity $N$ in ultracentral relativistic heavy-ion collisions: $\langle p_{T} \rangle \propto N^{b_{\rm UC}}$. It has been put forward that this exponent $b_{\rm UC}$ is the speed of s

Tobias Wolfgruber, Tobias Gesser, Marco Knöll, Pieter Maris

We perform a precision study of radii in Boron isotopes for multiple realistic interactions from chiral effective field theory. We obtain predictions of radii with combined many-body and interaction uncertainty quantification from ab initio no-core shell model calculations together with machine learning extrapolation methods. An extension to radius differenc

Horia D. Cornean, Massimo Moscolari

The gauge covariant magnetic perturbation theory is tailored for one-body Schr\"odinger operators perturbed by long-range magnetic fields. In this work we present a self-contained exposition of the method, by outlining its technical foundations and discussing the physical heuristics behind the proofs. We apply it in order to prove the stability of spectral g

Kang An, Yuxing Liu, Rui Pan, Yi Ren

Training deep neural networks is a structured optimization problem, because the parameters are naturally represented by matrices and tensors rather than by vectors. Under this structural representation, it has been widely observed that gradients are low-rank and Hessians are approximately block diagonal. These structured properties are crucial for designing

Zhuqing Wang, Ruochen Gao, Xiaoling Wu, Berislav Buča

Driven-dissipative many-body system supports nontrivial quantum phases absent in equilibrium. As a prominent example, the interplay between coherent driving and collective dissipation can lead to a dynamical quantum phase that spontaneously breaks time-translation symmetry. This so-called boundary time crystal (BTC) is fragile in the presence of local dissip

Alexei Ilyin, Sergey Zelik

The Voight regularization of the Navier--Stokes system is studied in a bounded domain and on the torus. In the 3D case we obtain new explicit bounds for the attractor dimension improving the previously known results. In the 2D case we show that the estimates so obtained converge to the known estimates for the attractor of the Navier--Stokes system as the reg

Jeremy Kahn, Zhenghao Rao

We prove the existence of surface subgroups within any cocompact lattice $\Gamma$ in $\mathrm{SO}(2n,1)$ for $n\geq2$. This result addresses the cases missing from the work of Hamenst\"adt in 2015, who constructed surface subgroups in cocompact lattices for all other rank-one semisimple Lie groups of non-compact type.

Yunhai Hu, Yilun Zhao, Chen Zhao, Arman Cohan

We introduce MCTS-RAG, a novel approach that enhances the reasoning capabilities of small language models on knowledge-intensive tasks by leveraging retrieval-augmented generation (RAG) to provide relevant context and Monte Carlo Tree Search (MCTS) to refine reasoning paths. MCTS-RAG dynamically integrates retrieval and reasoning through an iterative decisio

Chenxi Wang, Jizhan Fang, Xiang Chen, Bozhong Tian

Recent advancements in Large Multimodal Models (LMMs) have shown promise in Autonomous Driving Systems (ADS). However, their direct application to ADS is hindered by challenges such as misunderstanding of traffic knowledge, complex road conditions, and diverse states of vehicle. To address these challenges, we propose the use of Knowledge Editing, which enab

Marco Cattaneo, Louan Presse, Outi Supponen

Wall-attached bubbles can produce repeated jets under gentle ultrasound stimulation through the Faraday instability. We identify three distinct jetting regimes defined by the jetting frequency and the bubble surface topology. We demonstrate that these jets form via flow-focusing singularities following two distinct collapse modes of the bubble interface: con

Rianna Jitosho, Crystal E. Winston, Shengan Yang, Jinxin Li

Aerial robotic arms aim to enable inspection and environment interaction in otherwise hard-to-reach areas from the air. However, many aerial manipulators feature bulky or heavy robot manipulators mounted to large, high-payload aerial vehicles. Instead, we propose an aerial robotic arm with low mass and a small stowed configuration called a "flying vine". The

Fernando Torales Acosta, Tanvi Wamorkar, Vinicius Mikuni, Benjamin Nachman

Likelihood ratios are used for a variety of applications in particle physics data analysis, including parameter estimation, unfolding, and anomaly detection. When the data are high-dimensional, neural networks provide an effective tools for approximating these ratios. However, neural network training has an inherent stochasticity that limits their precision.

Huajie Tan, Yuheng Ji, Xiaoshuai Hao, Xiansheng Chen

Visual reasoning abilities play a crucial role in understanding complex multimodal data, advancing both domain-specific applications and artificial general intelligence (AGI). Existing methods enhance Vision-Language Models (VLMs) through Chain-of-Thought (CoT) supervised fine-tuning using meticulously annotated data. However, this approach may lead to overf

Mohammad Akyash, Kimia Zamiri Azar, Hadi Mardani Kamali

Grading programming assignments is a labor-intensive and time-consuming process that demands careful evaluation across multiple dimensions of the code. To overcome these challenges, automated grading systems are leveraged to enhance efficiency and reduce the workload on educators. Traditional automated grading systems often focus solely on correctness, faili

Leila Seidabadi, Indra Vandenbussche, Rowan Carter Fink, MacKenzie Moore

Objective: In this review, we aim to provide an analysis of current cardiac ablation techniques, such as radiofrequency ablation (RF), cryoablation, and pulsed-field ablation (PFA), with a focus on the role of computational modeling in enhancing the precision, safety, and effectiveness of these treatments. Particular attention is given to thermal management,

Soham Sane

This paper introduces a theoretical framework for a Transformer-augmented, sectional Mixture-of-Experts (MoE) architecture that aims to enhance computational efficiency while preserving model scalability. Unlike conventional MoE models, which route entire token embeddings to selected experts, our approach portions the embedding dimension itself -- assigning

Yuxuan Lu, Jing Huang, Yan Han, Bingsheng Yao

Recent research shows that LLM Agents can generate ``believable'' human behaviors via prompt-only methods, and such agents have been increasingly adopted in downstream applications. However, existing evaluation of these agents only focuses on qualitative believability (whether human raters think they are accurate), leaving open questions of whether LLM agent

Chen Tang, Xinzhu Ma, Encheng Su, Xiufeng Song

Traditional spatiotemporal models generally rely on task-specific architectures, which limit their generalizability and scalability across diverse tasks due to domain-specific design requirements. In this paper, we introduce \textbf{UniSTD}, a unified Transformer-based framework for spatiotemporal modeling, which is inspired by advances in recent foundation

Anna Anop, Aleksandr Murach

We build a solvability theory of elliptic boundary-value problems in normed Sobolev spaces of generalized smoothness for any integrability exponent $p>1$. The smoothness is given by a number parameter and a supplementary function parameter that varies slowly at infinity. These spaces are obtained by a combination of the methods of the complex interpolation w

Qing Yao, Kanishka Misra, Leonie Weissweiler, Kyle Mahowald

Language models (LMs) tend to show human-like preferences on a number of syntactic phenomena, but the extent to which these are attributable to direct exposure to the phenomena or more general properties of language is unclear. We explore this with the English dative alternation (DO: "gave Y the X" vs. PO: "gave the X to Y"), using a controlled rearing parad

Boyuan Chen, Hanxiao Jiang, Shaowei Liu, Saurabh Gupta

Envisioning physically plausible outcomes from a single image requires a deep understanding of the world's dynamics. To address this, we introduce PhysGen3D, a novel framework that transforms a single image into an amodal, camera-centric, interactive 3D scene. By combining advanced image-based geometric and semantic understanding with physics-based simulatio

Yanpeng Sun, Shan Zhang, Wei Tang, Aotian Chen

Diagrams represent a form of visual language that encodes abstract concepts and relationships through structured symbols and their spatial arrangements. Unlike natural images, they are inherently symbolic, and entirely artificial. They thus pose unique challenges for Multimodal Large Language Models (MLLMs) distinct from natural image processing. Recent stud

Guoqiang Zhang, Kenta Niwa, J. P. Lewis, Cedric Mesnage

We introduce relative and absolute position matching (RAPM), a diffusion distillation method resulting in high quality generation that can be trained efficiently on a single GPU. Recent diffusion distillation research has achieved excellent results for high-resolution text-to-image generation with methods such as phased consistency models (PCM) and improved

Joshua Dorrington, Sushovan Majhi, Atish Mitra, James Moukheiber

This paper explores the use of Topological Data Analysis (TDA) to investigate patterns in zonal-mean zonal winds of the Arctic, which make up the polar vortex, in order to better explain polar vortex dynamics. We demonstrate how TDA reveals significant topological features in this polar vortex data, and how they may relate these features to the collapse of t

Z. Zarezadeh, N. Zarezadeh

Despite the many challenges in exploratory data analysis, artificial neural networks have motivated strong interests in scientists and researchers both in theoretical as well as practical applications. Among sources of such popularity of artificial neural networks the ability of modeling non-linear dynamical systems, generalization, and adaptation possibilit

Peter Achim, Kemal Ozbek

We develop a novel framework for costly information acquisition in which a decision-maker learns about an unobserved state by choosing a signal distribution, with the cost of information determined by the distribution of noise in the signal. We show that a natural set of axioms admits a unique integral representation of the cost function, and we establish th

Alfred Mallet, Stefan Eriksson, Marc Swisdak, James Juno

We develop a new scaling theory for the resistive tearing mode instability of a current sheet with a strong shear flow across the layer. The growth rate decreases with increasing flow shear and is completely stabilized as the shear flow becomes Alfv\'enic: both in the constant-$\Psi$ regime, as in previous results, but we also show that the growth rate is in

Swetha Kambham, Hubert Jhonson, Sai Prathap Reddy Kambham

As artificial intelligence becomes more and more ingrained in daily life, we present a novel system that uses deep learning for music recommendation and emotion-based detection. Through the use of facial recognition and the DeepFace framework, our method analyses human emotions in real-time and then plays music that reflects the mood it has discovered. The s

Rajesh Mondal, Sourav Mondal, Toshifumi Yamada

In this paper, we investigate the production of Majorana fermionic dark matter (DM) via the Higgs portal, considering both freeze-in and freeze-out mechanisms during and after the post-inflationary reheating phase. We assume that the Universe is reheated through the decay of the inflaton ($\phi$) into a pair of fermions $f$ and $\bar f$ via the interaction $

Jeffery L Painter, François Haguinet, Gregory E Powell, Andrew Bate

Semantic similarity measures (SSMs) are widely used in biomedical research but remain underutilized in pharmacovigilance. This study evaluates six ontology-based SSMs for clustering MedDRA Preferred Terms (PTs) in drug safety data. Using the Unified Medical Language System (UMLS), we assess each method's ability to group PTs around medically meaningful centr

Chunhao Cai, Yiwu Shang

This paper introduces a new periodic fractional autoregressive process (PFAR) driven by fractional Gaussian noise (fGn) to model time series of precipitation evapotranspiration. Compared with the similar model in [\emph{Water Resources Research}, \textbf{20} (1984) 1898--1908], the new model incorporates a periodic structure via specialized varying coefficie

Max Carter

Let $G$ be a locally elliptic group, $(\Phi,\Psi)$ a complementary pair of Young functions, and $\omega: G \rightarrow [1,\infty)$ a weight function on $G$ such that the weighted Orlicz space $L^\Phi(G,\omega)$ is a Banach $*$-algebra when equipped with the convolution product and involution $f^*(x):=\overline{f(x^{-1})}$ ($f \in L^\Phi(G,\omega)$). Such a w

Ziyu Zhou, Keyan Hu, Yutian Fang, Xiaoping Rui

Change detection is a key task in Earth observation applications. Recently, deep learning methods have demonstrated strong performance and widespread application. However, change detection faces data scarcity due to the labor-intensive process of accurately aligning remote sensing images of the same area, which limits the performance of deep learning algorit

José Laurentino, Carlos Eduardo, Luiza Paffer, José Araújo

We report experiments that control magnetic damping in Py/Fe3O4-PANI via two spintronic thermal effects: spin Seebeck and anomalous Nernst. Magnetic damping is measured using ferromagnetic resonance (FMR) techniques, where the sample is excited by microwave radiation and the resulting DC voltage is detected in the Fe3O4-PANI film. When a temperature gradient

Gabriela A. Araujo, Laura Baudis, Nathaniel Bowden, Jordan Chapman

We present initial results on nuclear recoil detection based on the fluorescence of color centers created by nuclear recoils in lithium fluoride. We use gamma rays, fast and thermal neutrons, and study the difference in responses they induce, showing that this type of detector is rather insensitive to gamma rays. We use light-sheet fluorescence microscopy to

Masane Fuchi, Tomohiro Takagi

Score-based or diffusion models generate high-quality tabular data, surpassing GAN-based and VAE-based models. However, these methods require substantial training time. In this paper, we introduce RecTable, which uses the rectified flow modeling, applied in such as text-to-image generation and text-to-video generation. RecTable features a simple architecture

Juan Javier Diaz-Mejia, Elias Williams, Octavian Focsa, Dylan Mendonca

Many methods have been proposed for removing batch effects and aligning single-cell RNA (scRNA) datasets. However, performance is typically evaluated based on multiple parameters and few datasets, creating challenges in assessing which method is best for aligning data at scale. Here, we introduce the K-Neighbors Intersection (KNI) score, a single score that

Cheolhee Han, Nadav Katz, Eran Sela

We consider a generic quantum many-body system initiated at thermal equilibrium and driven by an external parameter, and discuss the prospect for measuring the work done by the varying parameter on the system. While existing methods are based on a full control of the system's Hamiltonian and are thus limited to few-level quantum systems, measuring work in ma

Joona V. Pankkonen, Lauri Ylinen, Matti Raasakka, Ilkka Tittonen

Variational quantum circuits (VQCs) are an essential tool in applying noisy intermediate-scale quantum computers to practical problems. VQCs are used as a central component in many algorithms, for example, in quantum machine learning, optimization, and quantum chemistry. Several methods have been developed to optimize VQCs. In this work, we enhance the perfo

Tohid Kargar Tasooji, Sakineh Khodadadi

This paper addresses the problem of distributed coordination control for multi-robot systems (MRSs) in the presence of localization uncertainty using a Linear Quadratic Gaussian (LQG) approach. We introduce a stochastic LQG control strategy that ensures the coordination of mobile robots while optimizing a performance criterion. The proposed control framework

Atsushi Ichino, Kartik Prasanna

For an odd prime $p$, we realize the trivial representation of $\mathrm{GL}_2(\mathbb{Z}/p^n\mathbb{Z})$ on the free $\mathbb{Z}/p^n \mathbb{Z}$-module of rank one as a subquotient of a direct sum of symmetric power representations (twisted by appropriate powers of the determinant) of rank strictly greater than one. The proof eventually reduces to establishi

Floriana Giannuzzi, Stefano Nicotri

We study the chiral condensate at finite temperature in AdS/QCD in a time-dependent background, in which the position of the black-hole horizon $z_h$ changes with time, producing an increasing or decreasing temperature. Conformal invariance is broken, as in the soft-wall model, by a static quadratic dilaton. Two different scenarios are analysed: in the first

Marko Petković, José-Manuel Vicent Luna, El\=ıza Beate Dinne, Vlado Menkovski

Accurately predicting adsorption properties in nanoporous materials using Deep Learning models remains a challenging task. This challenge becomes even more pronounced when attempting to generalize to structures that were not part of the training data.. In this work, we introduce SymGNN, a graph neural network architecture that leverages material symmetries t