November 2025 arXiv papers — page 120

Mingjie Xu, Jinpeng Chen, Yuzhi Zhao, Jason Chun Lok Li

Multimodal large language models (MLLMs) have enabled a wide range of advanced vision-language applications, including fine-grained object recognition and contextual understanding. When querying specific regions or objects in an image, human users naturally use "visual prompts" (VPs), such as bounding boxes, to provide reference. However, no existing benchma

Saad Mankarious, Ayah Zirikly, Daniel Wiechmann, Elma Kerz

Social media data has become a vital resource for studying mental health, offering real-time insights into thoughts, emotions, and behaviors that traditional methods often miss. Progress in this area has been facilitated by benchmark datasets for mental health analysis; however, most existing benchmarks have become outdated due to limited data availability,

Xuanyu Tian, Lixuan Chen, Qing Wu, Xiao Wang

Cardiac magnetic resonance (CMR) imaging is widely used to characterize cardiac morphology and function. To accelerate CMR imaging, various methods have been proposed to recover high-quality spatiotemporal CMR images from highly undersampled k-t space data. However, current CMR reconstruction techniques either fail to achieve satisfactory image quality or ar

Zhaoxin Chang, Shuguang Xiao, Fusang Zhang, Xujun Ma

Recently, Wi-Fi signals have emerged as a powerful tool for contactless sensing. During the sensing process, obtaining target direction information can provide valuable contextual insights for various applications. Existing direction estimation methods typically rely on antenna arrays, which are costly and complex to deploy in real-world scenarios. In this p

Sibgat Ul Islam, Jawad Ibn Ahad, Fuad Rahman, Mohammad Ruhul Amin

Knowledge Distillation (KD) trains a smaller student model using a large, pre-trained teacher model, with temperature as a key hyperparameter controlling the softness of output probabilities. Traditional methods use a fixed temperature throughout training, which is suboptimal. Moreover, architectural differences between teacher and student often result in mi

Maria-Teresa De Rosa Palmini, Eva Cetinic

The ambiguity between generalization and memorization in TTI diffusion models becomes pronounced when prompts invoke culturally shared visual references, a phenomenon we term multimodal iconicity. These are instances in which images and texts reflect established cultural associations, such as when a title recalls a familiar artwork or film scene. Such cases

Wei Chow, Jiachun Pan, Yongyuan Liang, Mingze Zhou

Recent advances in unified multimodal models (UMMs) have enabled impressive progress in visual comprehension and generation. However, existing datasets and benchmarks focus primarily on single-turn interactions, failing to capture the multi-turn, context-dependent nature of real-world image creation and editing. To address this gap, we present WEAVE, the fir

Giulio Grossi, Leo Vanciu, Veronica Ballerini, Danielle Braun

The harmful relationship between heatwaves and health has been extensively documented in medical and epidemiological literature. However, most evidence is associational and cannot be interpreted causally unless strong assumptions are made. In this paper, we first make explicit the assumptions underlying the statistical methods frequently used in the heatwave

Giada Peron

Massive Star Clusters (SCs) have been proposed as additional contributors to Galactic Cosmic rays (CRs), to overcome the limitations of supernova remnants (SNRs) to reach the highest energy end of the CR spectrum. Thanks to fast mass losses due to the collective stellar winds, the environment around SCs is potentially suitable for particle acceleration up to

Simon Badger, Matteo Becchetti, Colomba Brancaccio, Michal Czakon

The production of a top-quark pair, the heaviest known elementary particle, in association with a light jet is a key process for studying the properties of the Standard Model of Particle Physics. Due to its significance as a signal process with considerable sensitivity to the top-quark mass and as a background process for new physics searches, it is crucial

Jinbin Fu, Eric Laurendeau

Nonlinear unsteady vortex lattice-vortex particle methods (NL-UVLM-VPM) provide medium-fidelity predictions of rotorcraft aerodynamics with explicit three-dimensional wake representations at a moderate computational cost. This study presents an NL-UVLM-VPM approach with a scale-consistent adaptive wake panel-particle conversion strategy that mitigates the in

Lorenzo Ghiro, Marco Franceschini, Renato Lo Cigno, Michele Segata

The design of Cooperative Adaptive Cruise Control (CACC) algorithms for vehicle platooning has been extensively investigated, leading to a wide range of approaches with different requirements and performance. Most existing studies evaluate these algorithms under the assumption of homogeneous platoons, i.e., when all platoon members adopt the same CACC. Howev

Aymeric Delteil, Stéphanie Buil, Jean-Pierre Hermier

The potential of solid-state quantum emitters for applications critically depends on several key figures of merit. One of the most important is the quantum coherence of the emitted single photons, which can be compromised by fast dephasing and spectral diffusion. In hexagonal boron nitride (hBN), blue-emitting color centers (or B centers) are seen as favorab

Francisco Nogueira, Alexandre Bernardino, Bruno Martins

Referring Expression Comprehension (REC) requires models to localize objects in images based on different types of natural language descriptions. Even with significant progress, research on the area remains predominantly English-centric, despite increasing global deployment demands. This work addresses multilingual REC through two main contributions. First,

Vishnu Rajagopal, Puxun Wu

The non-commutative geometry offers an effective framework for describing physics at the Planck scale, incorporating generic quantum-gravitational effects through an intrinsic minimal length and the $\kappa$-deformed space-time stands out as a particularly well-developed model based on a Lie-algebraic type non-commutative space-time structure. We investigate

Cong-Tinh Dao, Nguyen Minh Thao Phan, Jun-En Ding, Chenwei Wu

Electronic health records (EHRs) are designed to synthesize diverse data types, including unstructured clinical notes, structured lab tests, and time-series visit data. Physicians draw on these multimodal and temporal sources of EHR data to form a comprehensive view of a patient's health, which is crucial for informed therapeutic decision-making. Yet, most p

Lukun Wu, Jie Li, Ziqi Ren, Kaifan Zhang

Decoding visual features from EEG signals is a central challenge in neuroscience, with cross-modal alignment as the dominant approach. We argue that the relationship between visual and brain modalities is fundamentally asymmetric, characterized by two critical gaps: a Fidelity Gap (stemming from EEG's inherent noise and signal degradation, vs. vision's high-

Dara Varam, Diaa A. Abuhani, Imran Zualkernan, Raghad AlDamani

Flow Matching (FM) generative models offer efficient simulation-free training and deterministic sampling, but their practical deployment is challenged by high-precision parameter requirements. We adapt optimal transport (OT)-based post-training quantization to FM models, minimizing the 2-Wasserstein distance between quantized and original weights, and system

Alessandro Bosso, Marco Borghesi, Andrea Iannelli, Bowen Yi

We present an approach to compute stabilizing controllers for continuous-time linear time-invariant systems directly from an input-output trajectory affected by process and measurement noise. The proposed output-feedback design combines (i) an observer of a non-minimal realization of the plant and (ii) a feedback law obtained from a linear matrix inequality

Sylvester Willys Namagwa

The efficiency of pension schemes in Kenya invites elevated interest owing to the increasing pension contribution amounts and the expectation that benefits paid out of these schemes would protect members from old age poverty. The study investigates the intervening effect of risk management on the relationship between corporate governance and the efficiency o

Nikolas Borrel-Jensen, Josiah Bjorgaard

We demonstrate a practical differentiable programming approach for acoustic inverse problems through two applications: admittance estimation and shape optimization for resonance damping. First, we show that JAX-FEM's automatic differentiation (AD) enables direct gradient-based estimation of complex boundary admittance from sparse pressure measurements, achie

Giulia Mariani, Chiara Lambranzi, Nicholas Cartocci, Giacinto Barresi

Lower-limb exosuits are particularly relevant for individuals with some degree of mobility impairment, such as post-stroke patients or older adults with reduced movement capabilities. This study aims to investigate the mental workload (MWL) assessment of XoSoft, a lower-limb soft exoskeleton, using and comparing subjective and objective physiological metrics

Wu-Fei Zhang

The Adaptive Exponential Integrate-and-Fire (AdEx) model is a simplified framework that effectively characterizes neuronal electrical activity. The aim of this paper is to employ phase plane analysis to systematically investigate diverse firing patterns generated by the AdEx model under varying parametric conditions. We first introduce the fundamental equati

Antoine Mazières, Thierry Poibeau

This data paper introduces MajinBook, an open catalogue designed to facilitate the use of shadow libraries-such as Library Genesis and Z-Library-for computational social science and cultural analytics. By linking metadata from these vast, crowd-sourced archives with structured bibliographic data from Goodreads, we create a high-precision corpus of over 539,0

Xingshuang Lin, Binbin Zhao, Jinwen Wang, Qinge Xie

Smart Contract Reusable Components(SCRs) play a vital role in accelerating the development of business-specific contracts by promoting modularity and code reuse. However, the risks associated with SCR usage violations have become a growing concern. One particular type of SCR usage violation, known as a logic-level usage violation, is becoming especially harm

Jiaxi Huang, Dongxu Wu, Hanwei Zhu, Lingyu Zhu

The rapid advancement of Multi-modal Large Language Models (MLLMs) has expanded their capabilities beyond high-level vision tasks. Nevertheless, their potential for Document Image Quality Assessment (DIQA) remains underexplored. To bridge this gap, we propose Q-Doc, a three-tiered evaluation framework for systematically probing DIQA capabilities of MLLMs at

Cyril Houdayer, Amine Marrakchi

We introduce the notion of selfless W$^*$-probability space and study its connection with Connes' bicentralizer problem. In particular, we show that if $M$ is a separable type ${\rm III_1}$ factor with trivial bicentralizer, then $(M, \varphi)$ is selfless for every faithful normal state $\varphi \in M_\ast$.

Yijie Chang, Shihang Tang, Haowen Deng, Yefeng Wang

Our early Universe is not perfectly homogeneous and it may contain some inhomogeneous sources, which might distort the local spacetime and modify the bubble nucleation rate. Taking the primordial black hole as an extreme example, we investigate the bubble nucleation rate of a first-order phase transition in the vicinity of primordial black holes or other pri

Manyu Li, Ruian He, Chenxi Ma, Weimin Tan

Multimodal Large Language Models are increasingly applied to biomedical imaging, yet scientific reasoning for microscopy remains limited by the scarcity of large-scale, high-quality training data. We introduce MicroVQA++, a three-stage, large-scale and high-quality microscopy VQA corpus derived from the BIOMEDICA archive. Stage one bootstraps supervision fro

Jianhao Su, Yanliang Zhang

The information released to investors in financial markets has various forms. We refer to range information as information about the upper and lower bound which the payoff of a risky asset may reach in the future. This study develops rational expectation models to explore the market impacts of disclosure of range information. Our model shows that its disclos

C. Hainaut, K. Ouahrouche, A. Rancon, G. Patera

Ultrafast imaging of photon statistics in two dimensions is a powerful tool for probing non-equilibrium and transient optical phenomena, yet it remains experimentally challenging due to the simultaneous need for high temporal resolution and statistical fidelity. In this work, we demonstrate spatially resolved single-shot measurements of photon number distrib

Amit Jain, Victor Rodriguez-Fernandez, Richard Linares

Autonomous spacecraft control for mission phases such as launch, ascent, stage separation, and orbit insertion remains a critical challenge due to the need for adaptive policies that generalize across dynamically distinct regimes. While reinforcement learning (RL) has shown promise in individual astrodynamics tasks, existing approaches often require separate

Shumit A. Mitra, Alexander Deisting, Jan Lommler, Uwe Oberlack

Next-generation dual-phase time projection chambers (TPCs) for rare event searches will require large-scale, high-precision electrodes. To meet the stringent requirements for high-voltage performance of such an experiment, we have developed a scanning setup for comprehensive electrode quality assurance. The system is built around the GRANITE (Granular Roboti

Alexander Deisting, Jan Lommler, Shumit A. Mitra, Uwe Oberlack

Next-generation dual-phase time projection chambers (TPCs) for rare event searches will require large-scale, high-precision electrodes. To meet the stringent requirements for mechanical stability and high-voltage performance of such an experiment, we have developed a scanning setup for electrode quality assurance called GRANITE: Granular Robotic Assay for No

Rosario Napoli, Antonio Celesti, Massimo Villari, Maria Fazio

Graph Machine Learning (GML) with Graph Databases (GDBs) has gained significant relevance in recent years, due to its ability to handle complex interconnected data and apply ML techniques using Graph Data Science (GDS). However, a critical gap exists in the current way GDB-GML applications analyze data, especially in terms of Knowledge Completion (KC) in Kno

Goryuchkina Irina

A Maillet-Malgrange type theorem is proved for a Dulac series (in the general case, with complex exponents), which formally satisfies an analytical ordinary differential equation (ODE). This theorem allows to estimate the growth of the norms of the coefficients of such a series, that is, to determine its Gevrey order, and in the special case it provides a su

Vincenzo Lipardi, Xenofon Chiotopoulos, Jacco A. de Vries, Domenica Dibenedetto

Quantum Computing is a rapidly developing field with the potential to tackle the increasing computational challenges faced in high-energy physics. In this work, we explore the potential and limitations of variational quantum algorithms in solving the particle track reconstruction problem. We present an analysis of two distinct formulations for identifying st

Chen Deng, Yong-Feng Huang, Abdusattar Kurban, Jin-Jun Geng

Short gamma-ray bursts (GRBs) exhibiting a plateau phase provide valuable insights into the post-merger activity of their central engines. Although the physical origin of the plateau remains uncertain, the magnetar energy injection model offers a compelling explanation that reproduces the observed temporal and luminosity features. However, previous studies r

Xiaoran Li, Jie Qiu, Liang Hao, Shiyang Zou

The evolution of stimulated Brillouin scattering (SBS) driven by broadband lasers in high-Z plasmas is investigated using one-dimensional collisional particle-in-cell simulations. The temporal incoherence of broadband lasers modulates the pump intensity, generating stochastic intensity pulses that intermittently drive SBS. The shortened coherence time weaken

Zejiao Liu, Yi Li, Jiali Wang, Junqi Tu

Multi-agent reinforcement learning (MARL) has made significant strides in enabling coordinated behaviors among autonomous agents. However, most existing approaches assume that communication is instantaneous, reliable, and has unlimited bandwidth; these conditions are rarely met in real-world deployments. This survey systematically reviews recent advances in

Owen A. Maute, Blake A. Roberts, Berker Peköz

This paper presents RadAround, a passive 2-D direction-finding system designed for adversarial IoT sensing in contested environments. Using mechanically steered narrow-beam antennas and field-deployable SCADA software, it generates high-resolution electromagnetic (EM) heatmaps using low-cost COTS or 3D-printed components. The microcontroller-deployable SCADA

Yeyue Cai, Jianhua Mo, Meixia Tao

Phase-time arrays, which integrate phase shifters (PSs) and true-time delays (TTDs), have emerged as a cost-effective architecture for generating frequency-dependent rainbow beams in wideband sensing and localization. This paper proposes an end-to-end deep learning-based scheme that simultaneously designs the rainbow beams and estimates user positions. Treat

Bernd Finkbeiner, Niklas Metzger, Satya Prakash Nayak, Anne-Kathrin Schmuck

\emph{Universal Safety Controllers (USCs)} are a promising logical control framework that guarantees the satisfaction of a given temporal safety specification when applied to any realizable plant model. Unlike traditional methods, which synthesize one logical controller over a given detailed plant model, USC synthesis constructs a \emph{generic controller} w

Jeffrey S. Bowers, Jeff Mitchell

According to Futrell and Mahowald [arXiv:2501.17047], both infants and language models (LMs) find attested languages easier to learn than impossible languages that have unnatural structures. We review the literature and show that LMs often learn attested and many impossible languages equally well. Difficult to learn impossible languages are simply more compl

Sanath Keshav, Felix Fritzen

We propose a spectrally normalized surrogate for forward and inverse mechanical homogenization with hard physical guarantees. Leveraging the Voigt-Reuss bounds, we factor their difference via a Cholesky-like operator and learn a dimensionless, symmetric positive semi-definite representation with eigenvalues in $[0,1]$; the inverse map returns symmetric posit

Peter Mader

Davydov-Yetter (DY) cohomology is a cohomology theory for linear semigroupal (i.e.~monoidal but not necessarily categories and functors, measuring deformations of their coherence isomorphisms. We show that DY cohomology is invariant under freely adjoining a unit object, and under adjoining colimits. This implies that constructions such as Ind-completion and

Junzhe Song, Ruisi He, Mi Yang, Zhengyu Zhang

With the rapid deployments of 5G and 6G networks, accurate modeling of urban radio propagation has become critical for system design and network planning. However, conventional statistical or empirical models fail to fully capture the influence of detailed geometric features on site-specific channel variances in dense urban environments. In this paper, we pr

Nguyen Xuan Duy Bao, Nguyen Mau Nam

We revisit classical gradient characterizations of quasiconvexity and provide corrected proofs that close gaps in earlier arguments. For the differentiable case of $\sigma$-quasiconvexity, we establish the full equivalence between several first-order conditions, resolving a remaining implication left open in the recent literature. Our approach yields a conci

Tim J. Boonen, Engel John C. Dela Vega

This paper considers an insurer with two collaborating business lines that must make three critical decisions: (1) dividend payout, (2) a combination of proportional and excess-of-loss reinsurance coverage, and (3) capital injection between the lines. The reserve level of each line is modeled using a diffusion approximation, with the insurer's objective bein

Gonzalo Barriga, Carla Henríquez-Baez, Leonardo Sanhueza, Aldo Vera

In this paper, we construct self-gravitating topological solitons in the $SU(N)$ Einstein non-linear sigma model coupled to $\omega$-vector mesons in four space-time dimensions. These solutions represent tube-like configurations free of curvature singularities, carrying a non-vanishing topological charge that is identified as the baryon number. We show that

Gioliano de Oliveira Braga, Pedro Henrique dos Santos Rocha, Rafael Pimenta de Mattos Paixão, Giovani Hoff da Costa

Wi-Fi Channel State Information (CSI) has been repeatedly proposed as a biometric modality, often with reports of high accuracy and operational feasibility. However, the field lacks a consolidated understanding of its security properties, adversarial resilience, and methodological consistency. This Systematization of Knowledge (SoK) examines CSI-based biomet

Jiangkai Long, Yanran Zhu, Chang Tang, Kun Sun

Spatial transcriptomics enables gene expression profiling with spatial context, offering unprecedented insights into the tissue microenvironment. However, most computational models treat genes as isolated numerical features, ignoring the rich biological semantics encoded in their symbols. This prevents a truly deep understanding of critical biological charac

Davi de Moura Esposti Moreira, Matheus Elias Pereira, Alexandre Grezzi de Miranda Schmidt

Tachyons are hypothetical particles that propagate faster than light, yet they have never been observed in nature or in the laboratory. In this work, we introduce the hydrogen molecule ion as an analogue for the dynamics of a spinless test particle interacting with the gravitational field generated by a tachyon. The tachyonic spacetime is modeled using an AI

Julian Grolig, Lars Griem, Michael Selzer, Hans-Ulrich Kauczor

Extracting digital material representations from images is a necessary prerequisite for a quantitative analysis of material properties. Different segmentation approaches have been extensively studied in the past to achieve this task, but were often lacking accuracy or speed. With the advent of machine learning, supervised convolutional neural networks (CNNs)

Erik Rosenberg, Jens Chluba

We extend the recent boost operator formalism for relativistic Compton scattering calculations to also account for polarization. This allows us to provide general, exact expressions for the polarized Sunyaev-Zeldovich (SZ) effect sourced both kinematically and from intrinsic anisotropies of the Cosmic Microwave Background (CMB). The results are given in term

Francesco Bisio

For $\Bbbk$ a field, let $X$ a $m \times n$ matrix of variables and $S=\Bbbk[X].$ We consider the determinantal ideal $I_2 \subseteq S$ generated by the $2$-minors of $X.$ In this paper we find a suitable monomial order over $S$ such that $I$, the initial ideal of $I_2$ with respect to that order, is level, namely, it is Cohen-Macaulay and the socle of an Ar

Stephen Gould, Tom Kelly

Let $H$ be a $(k+1)$-uniform hypergraph which is nearly $D$-regular, such that any set of $i$ vertices is contained in at most $D_i$ edges of $H$ for each $i = 2, 3, \dots, k+1$. Influential results of Pippenger and of Frankl and R\"odl show that the \textit{R\"odl Nibble} -- a probabilistic procedure which iteratively constructs a matching in small bits --

Nicola Piovella, Romain Bachelard

Subradiance is a phenomenon where coupled emitters radiate light at a slower rate than independent ones. While its observation was first reported in disordered cold atom clouds, ordered subwavelength arrays of emitters have emerged as promising platforms to design highly cooperative optical properties based on dipolar interactions. In this work we characteri

Kaizheng Wang, Fabio Cuzzolin, David Moens, Hans Hallez

Deep ensembles (DE) have emerged as a powerful approach for quantifying predictive uncertainty and distinguishing its aleatoric and epistemic components, thereby enhancing model robustness and reliability. However, their high computational and memory costs during inference pose significant challenges for wide practical deployment. To overcome this issue, we

Shulin Liu, Dong Du, Tao Yang, Yang Li

Recent progress in large language models (LLMs) has been propelled by reinforcement learning with verifiable rewards (RLVR) and test-time scaling. However, the limited output length of LLMs constrains the depth of reasoning attainable in a single inference process. Multi-agent reasoning systems offer a promising alternative by employing multiple agents inclu

Luke Alexander Turnbull, Max Thomas Birch, Marisel Di Pietro Martínez, Rikako Yamamoto

Nanoscale topologically non-trivial magnetization configurations generate significant interest due to both the fundamental properties of their knotted structures and their potential applications in ultra-efficient computing devices. While such textures have been widely studied in two dimensions, three-dimensional (3D) systems can yield more complex configura

Daniel Ebert, Antonia Ellerbrock

We study the recently introduced fair division concept of the happy nucleolus for cost allocation among players in a cooperative game, with special focus on its computation. The happy nucleolus applies the same fairness criterion as the well-established nucleolus but with reduced total value. Still, we show that the relation between the two concepts is quite

Jiahao Wang, Bokang Fu, Yu Zhu, Yuli Liu

LLM-based agents are emerging as a promising paradigm for simulating user behavior to enhance recommender systems. However, their effectiveness is often limited by existing studies that focus on modeling user ratings for individual items. This point-wise approach leads to prevalent issues such as inaccurate user preference comprehension and rigid item-semant

Adeleh Mazaherian, Erfan Nourbakhsh

The rapid integration of generative artificial intelligence (AI) in higher education since 2023 has outpaced institutional preparedness, creating a persistent gap between student practices and established ethical standards. This paper draws on mixed-method surveys and a focused literature review to examine student motivations, ethical dilemmas, gendered resp

Sheng Liu, Yuanzhi Liang, Sidan Du

Recent 3D human motion generation models demonstrate remarkable reconstruction accuracy yet struggle to generalize beyond training distributions. This limitation arises partly from the use of precise 3D supervision, which encourages models to fit fixed coordinate patterns instead of learning the essential 3D structure and motion semantic cues required for ro

Kristjan Kannike, Luca Marzola, Kristjan Müürsepp

Multi-phase critical scenarios explain the observed Higgs boson mass scale by the almost simultaneous occurrence of two smoothly connected phases of the theory, which differ by the selected vacuum configuration. A generic prediction of the framework is the presence of a further light scalar state, the dilaton, which naturally couples weakly to the Higgs boso

Benjamin Blakely, Daniel Karcz

Modern electrical power grids represent complex cyber-physical systems requiring specialized cybersecurity frameworks beyond traditional IT security models. Existing threat intelligence standards such as STIX 2.1 and MITRE ATT\&CK lack coverage for grid-specific assets, operational technology relationships, and cyber-physical interdependencies essential for

Piotr Faliszewski, Stanislaw Kazmierowski, Grzegorz Lisowski, Ildiko Schlotter

We study strategic candidate nomination by parties in elections decided by Plurality voting. Each party selects a nominee before the election, and the winner is chosen from the nominated candidates based on the voters' preferences. We introduce a new restriction on these preferences, which we call party-aligned single-peakedness: all voters agree on a common

Pomazanov Mikhail

The paper shows how to determine the loss on an LGD borrower's loan after default, with or without preparation of a separate model. LGD after default is estimated taking into account the average repayment period of the defaulted loan, knowledge of volumes, moments of default and repayments, the rate or other parameters in the vector of determinants. The calc

Giulio Schimperna, Antonio Segatti

We consider a mathematical model coupling the Cahn-Hilliard system for phase separation with an additional equation describing the diffusion process of a chemical quantity whose concentration influences the physical process. The main application of the model refers to tumor progression, where the phase variable denotes the local proportion of active cancer c

Prabodh Katti, Houssem Sifaou, Sangwoo Park, Bipin Rajendran

On-device fine-tuning is a critical capability for edge AI systems, which must support adaptation to different agentic tasks under stringent memory constraints. Conventional backpropagation (BP)-based training requires storing layer activations and optimizer states, a demand that can be only partially alleviated through checkpointing. In edge deployments in

Guangyi Dong, Zhihui Wang

Machine learning force fields (MLFFs), which employ neural networks to map atomic structures to system energies, effectively combine the high accuracy of first-principles calculation with the computational efficiency of empirical force fields. They are widely used in computational materials simulations. However, the development and application of MLFFs for l

Xuming Li, Cheng Peng, Chunhua Zhang, Xinnan Wu

Phase field lattice Boltzmann (LB) models have undergone continuous development, resulting in multiple variants widely used for simulating multiphase flows. However, direct performance comparisons remain limited, especially for three-dimensional cases. In this study, we present a systematic comparative analysis of several recent and representative phase-fiel

Giulio Antonio Abbo, Tony Belpaeme

Large language models are increasingly used in applications where alignment with human values is critical. While model fine-tuning is often employed to ensure safe responses, this technique is static and does not lend itself to everyday situations involving dynamic values and preferences. In this paper, we present a practical, reproducible, and model-agnosti

Matthew X. Burns, Jiaming Liang

We propose Log-Averaged Mirror Prox (LAMP), a linear-space primal-dual method for large-scale optimal transport. LAMP implements primal mirror prox updates by tracking an averaged dual sequence, reducing storage complexity from ${O}(nm)$ to $O(n+m)$ while preserving dense, GPU-friendly reductions. Consequently, LAMP preserves the last-iterate $\widetilde{O}(

Shangjia Zhang, Zhaohuan Zhu, Callum W. Fairbairn

Shadows are commonly observed in protoplanetary disks in near-infrared and (sub)millimeter images, often cast by misaligned inner disks or other obscuring material. While recent studies show that shadows can alter disk dynamics, only the case symmetric across the midplane (e.g., from a polar-aligned inner disk) has been studied. Here we study shadows cast by

Haixin Li, Yanke Li, Diego Paez-Granados

We introduce KarmaTS, an interactive framework for constructing lag-indexed, executable spatiotemporal causal graphical models for multivariate time series (MTS) simulation. Motivated by the challenge of access-restricted physiological data, KarmaTS generates synthetic MTS with known causal dynamics and augments real-world datasets with expert knowledge. The

Yanbo Dai, Zongjie Li, Zhenlan Ji, Shuai Wang

Large language models (LLMs) have achieved remarkable success across a wide range of natural language processing tasks, demonstrating human-level performance in text generation, reasoning, and question answering. However, training such models requires substantial computational resources, large curated datasets, and sophisticated alignment procedures. As a re

Lama Sleem, Jerome Francois, Lujun Li, Nathan Foucher

Jailbreak attacks designed to bypass safety mechanisms pose a serious threat by prompting LLMs to generate harmful or inappropriate content, despite alignment with ethical guidelines. Crafting universal filtering rules remains difficult due to their inherent dependence on specific contexts. To address these challenges without relying on threshold calibration

Ryan Cecil, Lucas Mentch

Classical model selection seeks to find a single model within a particular class that optimizes some pre-specified criteria, such as maximizing a likelihood or minimizing a risk. More recently, there has been an increased interest in model set selection (MSS), where the aim is to identify a (confidence) set of near-optimal models. Here, we generalize the MSS

Jiong-Hang Liang, Yunfeng Xiong

As a universal quantum mechanical approach to the dynamical many-body problem, the time-dependent density functional theory (TDDFT) might be inadequate to describe crucial observables that rely on two-body evolution behavior, like the double-excitation probability and two-body dynamic correlation. One promising remedy is to utilize the time-dependent 2-reduc

Jordan M. Adams, Daniel Heligman, Rajind Mendis, Josh Wetherington

In fluids, vortices form at boundaries between flows of different velocities. Here, we show that pulsed electromagnetic waves form spatiotemporal vortices when light straddles media of different phase velocities. When the resulting relative time-delay is on the order of the pulse duration of light, spatiotemporal optical vortices (STOVs) are formed. This met

Shengjie Sun, Jiafei Lyu, Runze Liu, Mengbei Yan

Offline imitation learning (offline IL) enables training effective policies without requiring explicit reward annotations. Recent approaches attempt to estimate rewards for unlabeled datasets using a small set of expert demonstrations. However, these methods often assume that the similarity between a trajectory and an expert demonstration is positively corre

Jason Beh, Jerome Morio, Florian Simatos

Motivated by the estimation of covariance matrices by importance sampling arising in the cross-entropy (CE) algorithm, we study a random matrix model $\hat \Sigma = {\bf X} L {\bf X}^\top$ with two distinct features: $\bf X$ and $L$ are dependent, and $L$ is heavy-tailed. In the high-dimensional regime $d \to \infty$, we prove under suitable assumptions that

Karl Kunisch, Jesper Schröder

Taking a deterministic viewpoint this work investigates extensions of the Kalman-Bucy filter for state reconstruction to systems containing parametric uncertainty in the state operator. The emphasis lies on risk averse designs reducing the probability of large reconstruction errors. In a theoretical analysis error bounds in terms of the variance of the uncer

Shin Kaneshiro, Robert Peters

The non-Hermitian skin effect (NHSE), characterized by the extensive localization of bulk modes at the boundaries, has attracted significant attention as a hallmark feature of non-Hermitian topology. This localization invalidates the conventional Bloch band theory, necessitating an analysis under open boundary conditions even in the thermodynamic limit. The

Christopher J. Fewster, Christiane K. M. Klein

The Proca field describes a massive relativistic spin-$1$ particle and was originally formulated in Minkowski spacetime. Here we consider a variety of generalizations in globally hyperbolic spacetimes, including couplings between a number of Proca fields via a mass-matrix, the charged Proca field with arbitrary magnetic moment in an arbitrary external electr

Shaowei Guan, Hin Chi Kwok, Ngai Fong Law, Gregor Stiglic

Retrieval-augmented generation (RAG) has rapidly emerged as a transformative approach for integrating large language models into clinical and biomedical workflows. However, privacy risks, such as protected health information (PHI) exposure, remain inconsistently mitigated. This review provides a thorough analysis of the current landscape of RAG applications

Matthew A. Morgan, Matt Bauwens, Seng Loo, Miho Hunter

We report on the development of transmission-line reflectionless filters operating with passbands at 100 GHz and 230 GHz, and stopband absorption up to 500 GHz, the highest operating frequencies yet recorded for such filters. The designs are based on a previously reported mathematical solution to the reflectionless condition, now successfully implemented for

Arun Thangamani, Md Asghar Ahmad Shahid, Adam Siemieniuk, Rolf Morel

The rapidly evolving landscape of AI and machine learning workloads has widened the gap between high-level domain operations and efficient hardware utilization. Achieving near-peak performance still demands deep hardware expertise-experts either handcraft target-specific kernels (e.g., DeepSeek) or rely on specialized libraries (e.g., CUTLASS)-both of which

Pavel Rojtberg, Julius Kühn

We introduce YCB-Ev SD, a synthetic dataset of event-camera data at standard definition (SD) resolution for 6DoF object pose estimation. While synthetic data has become fundamental in frame-based computer vision, event-based vision lacks comparable comprehensive resources. Addressing this gap, we present 50,000 event sequences of 34 ms duration each, synthes

Egemen Sert, Şeyda Ertekin

Multimodal reasoning has become a cornerstone of modern AI research. Standardized exam questions offer a uniquely rigorous testbed for such reasoning, providing structured visual contexts and verifiable answers. While recent progress has largely focused on algorithmic advances such as reinforcement learning (e.g., GRPO, DPO), the data centric foundations of

Marcello Baldo

The compatibility of special relativity and Quantum Mechanics has been questioned by several authors. The origin of this tension can be traced back mainly to the introduction of the measurement processes and the corresponding wave function reduction, which play a crucial role in Quantum Mechanics. We approach this problem with the help of a recent proposal f

E. Martello, G. A. Falci, E. Paladino, F. M. D. Pellegrino

We develop a Fokker-Planck approach to describe the dynamics of staggered magnetization and thermal fluctuations in a two-dimensional antiferromagnetic system with uniaxial anisotropy. Beginning with a classical model for the antiferromagnetic system, we incorporate a Landau-Lifshitz-Gilbert equation augmented by Langevin fields to account for thermal fluctu

Salima Lamsiyah, Saad Ezzini, Abdelkader El Mahdaouy, Hamza Alami

The generation of highly fluent text by Large Language Models (LLMs) poses a significant challenge to information integrity and academic research. In this paper, we introduce the Multi-Domain Detection of AI-Generated Text (M-DAIGT) shared task, which focuses on detecting AI-generated text across multiple domains, particularly in news articles and academic w

Davood Momeni

Recent analyses of low-redshift supernova and Cepheid data reveal localized shifts in the distance modulus, often interpreted as calibration anomalies or hints of new physics. We propose that these features may emerge naturally from environment-dependent modifications to gravity. In particular, we examine the Starobinsky \(f(R) = R + \lambda R^2\) model, whi

Stéphane Girard, Cambyse Pakzad

This paper develops a theoretical framework for Extreme Partial Least Squares (EPLS) dimension reduction in the presence of missing data and weak temporal dependence. Building upon the recent EPLS methodology for modeling extremal dependence between a response variable and high-dimensional covariates, we extend the approach to more realistic data settings wh

Akshat Gupta, Caterina Doglioni, Thomas Joseph Elliott

The petabyte-scale data generated annually by High Energy Physics (HEP) experiments like those at the Large Hadron Collider present a significant data storage challenge. Whilst traditional algorithms like LZMA and ZLIB are widely used, they often fail to exploit the deep structure inherent in scientific data. We investigate the application of modern state sp

Bhavnesh Bhat, Michael J. Keith, Ismaël Cognard, Lucas Guillemot

We report the detection of a glitch in the millisecond pulsar (MSP) PSR J0900$-$3144, which is included in the European, MeerKAT and Parkes pulsar timing array experiments. The dataset combines observations from the MeerKAT, Nan\c{c}ay, Lovell, and Murriyang telescopes, spanning a total baseline of approximately 14 years. The glitch occurred on MJD~59942(17)

Yarkın Gevez, Aymen Khaleel, Ertugrul Basar

This letter introduces a novel partitioning scheme for reconfigurable intelligent surfaces (RISs) that simultaneously consider RIS identification and beamforming. The proposed scheme dynamicly and efficiently allocates RIS elements between identification and beamforming users, considering the different performance metrics associated with each of them. By emp