October 2025 arXiv papers — page 50

Shireen Kudukkil Manchingal, Armand Amaritei, Mihir Gohad, Maryam Sultana

Autonomous Vehicle (AV) perception systems have advanced rapidly in recent years, providing vehicles with the ability to accurately interpret their environment. Perception systems remain susceptible to errors caused by overly-confident predictions in the case of rare events or out-of-sample data. This study equips an autonomous vehicle with the ability to 'k

Juan C. Leon Alcazar, Mattia Soldan, Mohammad Saatialsoruji, Alejandro Pardo

Compressed file formats are the corner stone of efficient data storage and transmission, yet their potential for representation learning remains largely underexplored. We introduce TEMPEST (TransformErs froM comPressed rEpreSenTations), a method that exploits the inherent byte-stream structure of compressed files to design an effective tokenization and encod

Yuval Kainan, Shaked Zychlinski

Large Language Models (LLMs) often expend significant computational resources generating boilerplate responses, such as refusals, simple acknowledgements and casual greetings, which adds unnecessary cost and latency. To address this inefficiency, we propose a simple yet highly effective method for detecting such responses after only a single generation step.

Dimitri Dine, Jack J Garzella

Let $K$ be a perfectoid field. We describe all quotient fields of the perfectoid Tate algebra\begin{equation*}T_{n,K}^{\text{perfd}}=K\langle X_{1}^{1/p^{\infty}},\dots, X_{n}^{1/p^{\infty}}\rangle\end{equation*}in any number $n\geq1$ of variables in terms of (completed perfections of) the nonarchimedean fields $K_{r_1,\dots,r_l}$ occuring in Berkovich geome

Netanel P. Yaish, Samata Gokhale, Avi Peer

We demonstrate both the generation and detection of an ultra-high flux of polarization Bell states using broadband hyper-entangled bi-photons that are quantum-correlated in both polarization and time-energy. Bell states of polarization embody the most basic form of two-state entanglement, and are a key component of quantum protocols of communication and sens

Laure Zanna, William Gregory, Pavel Perezhogin, Aakash Sane

Climate simulations, at all grid resolutions, rely on approximations that encapsulate the forcing due to unresolved processes on resolved variables, known as parameterizations. Parameterizations often lead to inaccuracies in climate models, with significant biases in the physics of key climate phenomena. Advances in artificial intelligence (AI) are now direc

Jinpeng Dong, Chen Li, Yutong Lin, Jingwen Fu

Predicting High-definition (HD) map elements with high quality (high classification and localization scores) is crucial to the safety of autonomous driving vehicles. However, current methods perform poorly in high quality predictions due to inherent task misalignment. Two main factors are responsible for misalignment: 1) inappropriate task labels due to one-

L. Hemanth Krishna, Srinivasu Bodapati, Sreehari Veeramachaneni, BhaskaraRao Jammu

This paper presents an approximate signed multiplier architecture that incorporates a sign-focused compressor, specifically designed for edge detection applications in machine learning and signal processing. The multiplier incorporates two types of sign-focused compressors: A + B + C + 1 and A + B + C + D + 1. Both exact and approximate compressor designs ar

Hagay Michaeli, Daniel Soudry

Transformers have emerged as a competitive alternative to convnets in vision tasks, yet they lack the architectural inductive bias of convnets, which may hinder their potential performance. Specifically, Vision Transformers (ViTs) are not translation-invariant and are more sensitive to minor image translations than standard convnets. Previous studies have sh

Anna Deichler, Jonas Beskow

We introduce Look and Tell, a multimodal dataset for studying referential communication across egocentric and exocentric perspectives. Using Meta Project Aria smart glasses and stationary cameras, we recorded synchronized gaze, speech, and video as 25 participants instructed a partner to identify ingredients in a kitchen. Combined with 3D scene reconstructio

Zhou Li, Siyan Qin, Xiang Zhang, Jihao Fan

In the Conditional Disclosure of Secrets (CDS) problem, Alice and Bob hold inputs $x\in \mathcal{X}$ and $y\in \mathcal{Y}$ and share a secret. Let $f:\mathcal{X}\times\mathcal{Y}\to\{0,1\}$ be a function such that the secret is revealed to a third party, Carol, if and only if $f(x,y)=1$. To protect the secret when $f(x,y)=0$, Alice and Bob share a common no

Xuan Lu, Haohang Huang, Rui Meng, Yaohui Jin

Large Language Models (LLMs) have recently demonstrated strong capabilities in tool use, yet progress in tool retrieval remains hindered by incomplete and heterogeneous tool documentation. To address this challenge, we introduce Tool-DE, a new benchmark and framework that systematically enriches tool documentation with structured fields to enable more effect

Wenkai Zhu, Xu Li, Qimin Xu, Benwu Wang

3D Gaussian Splatting SLAM has emerged as a widely used technique for high-fidelity mapping in spatial intelligence. However, existing methods often rely on a single representation scheme, which limits their performance in large-scale dynamic outdoor scenes and leads to cumulative pose errors and scale ambiguity. To address these challenges, we propose \text

D. K. He, Z. Song

We study the Hilbert space fragmentation (HSF) in hardcore Bose and Fermi Hubbard models in the framework of the restricted spectrum generating algebra (RSGA). We present a family of hardcore Bose-Hubbard models with repulsive density-density interactions on a generalized Lieb lattice. We show that this system possesses the RSGA structure in the large intera

Shunta Akiyama

In this paper, we consider a block coordinate descent (BCD) algorithm for training deep neural networks and provide a new global convergence guarantee under strictly monotonically increasing activation functions. While existing works demonstrate convergence to stationary points for BCD in neural networks, our contribution is the first to prove convergence to

Elias Koschier

The objective of this pre-scientific paper is to analyze the crisis management during the Fukushima nuclear accident in the year of 2011 and to derive lessons for a safer operation of reactors along with a more resilient crisis management framework. In addition to official reports and subject literature, the conducted methods included interviews with an IAEA

Tomoi Koide, Armin van de Venn

We show that when a Brownian bridge is physically constrained to satisfy a canonical condition, its time evolution exactly coincides with an m-geodesic on the statistical manifold of Gaussian distributions. This identification provides a direct physical realization of a geometric concept in information geometry. It implies that purely random processes evolve

Kazuyuki Yagasaki

We study bifurcations of twisted solutions in a continuum limit (CL) for the Kuramoto model (KM) of identical oscillators defined on nearest neighbor graphs, which may be deterministic dense, random dense or random sparse, when it may have phase-lag. We use the center manifold reduction, which is a standard technique in dynamical systems theory, and prove th

Bowie Liu, Dennis Wong, Chan-Tong Lam, Sio-Kei Im

We present the first known pivot Gray code for spanning trees of complete graphs, listing all spanning trees such that consecutive trees differ by pivoting a single edge around a vertex. This pivot Gray code thus addresses an open problem posed by Knuth in The Art of Computer Programming, Volume 4 (Exercise 101, Section 7.2.1.6, [Knuth, 2011]), rated at a di

Malik Imran, Safiullah Khan, Zain Ul Abideen, Ciara Rafferty

Post-quantum multivariate public key cryptography (MPKC) schemes resist quantum threats but require heavy operations, such as rejection sampling, which challenge resource-limited devices. Prior hardware designs have addressed various aspects of MPKC signature generation. However, rejection sampling remains largely unexplored in such contexts. This paper pres

Tristan Barkman

A new operator-level necessity result for the Chapman--Enskog expansion is established: in closed and unforced kinetic systems, the $O(\varepsilon)$ deviatoric stress arises if and only if the first Chapman--Enskog correction $f^{(1)}$ is nonzero. This resolves a gap in the classical kinetic-to-continuum literature, where the presence of first-order deviator

Panagiotis Gianniotis, Konstantinos Leskas

We show $L^1$-bounds of the Riemann curvature tensor on a smooth closed $n$-dimensional Ricci flow. To achieve this we introduce the notion of a neck of maximal symmetry, similar to the one in Cheeger-Jiang-Naber and Jiang-Naber and establish a decomposition result by balls with uniform curvature bounds that satisfy an appropriate $(n-2)$-content estimate.

Shu-Ning Ding, Ling-Fang Wang, Xiao-Tian Cheng, Jia-Wang Yu

The demand for high-efficiency and miniaturized on-chip light sources drives continuous innovation in photonic crystal (PhC) microcavity lasers. The presence of slow-light effects in PhC microcavities leads to the mode competition between Bloch modes resulting in multi-mode lasing, which obstructs the dense integration of PhC lasers. Here, we theoretically v

Josu Eguiluz Castañeira, Axel Brando, Migle Laukyte, Marc Serra-Vidal

Artificial intelligence (AI) now permeates critical infrastructures and decision-making systems where failures produce social, economic, and democratic harm. This position paper challenges the entrenched belief that regulation and innovation are opposites. As evidenced by analogies from aviation, pharmaceuticals, and welfare systems and recent cases of synth

Xiaoyan Li, Christoforos N. Hadjicostis

Attacks, including the manipulation of sensor readings and the modification of actuator commands, pose a significant challenge to the security and privacy of automated systems. This paper considers discrete event systems that can be modeled with nondeterministic finite state automata that are susceptible to state attacks. A state attack allows an intruder to

Zilong Wang, Qingtian Zeng, Hua Duan, Cheng Cheng

Few-shot Knowledge Graph Completion (FKGC) infers missing triples from limited support samples, tackling long-tail distribution challenges. Existing methods, however, struggle to capture complex relational patterns and mitigate data sparsity. To address these challenges, we propose a novel FKGC framework for conjugate relation modeling (CR-FKGC). Specificall

Berken Utku Demirel, Christian Holz

Self-supervised learning (SSL) has emerged as a powerful paradigm for learning representations without labeled data. Most SSL approaches rely on strong, well-established, handcrafted data augmentations to generate diverse views for representation learning. However, designing such augmentations requires domain-specific knowledge and implicitly imposes represe

Ilgam Latypov, Alexandra Suvorikova, Alexey Kroshnin, Alexander Gasnikov

In many modern applications, a system must dynamically choose between several adaptive learning algorithms that are trained online. Examples include model selection in streaming environments, switching between trading strategies in finance, and orchestrating multiple contextual bandit or reinforcement learning agents. At each round, a learner must select one

Maria Bevilacqua

We investigate models of algebraic theories in the category of cocommutative coalgebras over a field. We establish some of their categorical properties, similar to those of algebraic varieties. We introduce a class of categories of coalgebraic models of algebraic theories endowed with an underlying structure of cocommutative Hopf algebra, and show that these

Sofiane Ennadir, Johannes F. Lutzeyer, Michalis Vazirgiannis, El Houcine Bergou

Graph Neural Networks (GNNs) have demonstrated remarkable performance across a spectrum of graph-related tasks, however concerns persist regarding their vulnerability to adversarial perturbations. While prevailing defense strategies focus primarily on pre-processing techniques and adaptive message-passing schemes, this study delves into an under-explored dim

Lu Xu, Tsai Hor Chan, Kwok Fai Lam, Lequan Yu

Density estimation is essential for generative modeling, particularly with the rise of modern neural networks. While existing methods capture complex data distributions, they often lack interpretability and uncertainty quantification. Bayesian nonparametric methods, especially the \polya tree, offer a robust framework that addresses these issues by accuratel

Tharun Anand, Mohammad Hassan Vali, Arno Solin, Green Rosh

Diffusion models achieve remarkable fidelity in image synthesis, yet precise control over their outputs for targeted editing remains challenging. A key step toward controllability is to identify interpretable directions in the model's latent representations that correspond to semantic attributes. Existing approaches for finding interpretable directions typic

Inkyou Lee, Churlhi Lyi, Youngkuk Kim

Pt-Sr binary intermetallics encompass a broad range of stoichiometries and crystal structures, stabilized by complex bonding and multivalent chemistry. The Sr-rich end member, PtSr5, is recently identified via artificial-intelligence-guided materials design as a body-centered tetragonal compound (I4/m). Using first-principles calculations, we show that PtSr5

Xiu-Lei Ren, E. Epelbaum, J. Gegelia

Recently the nucleon-nucleon interaction derived using time-ordered perturbation theory in manifestly Lorentz-invariant chiral effective field theory was shown to yield promising results for peripheral neutron-proton scattering. In this work we study low partial waves at next-to-leading order by treating the potential non-perturbatively in the scattering equ

Nabajyoti Borah, Raju Moni Borah, Bandan Boruah, Purnendu Bikash Acharjee

The proposed solution is Deep Learning Technique that will be able classify three types of tea leaves diseases from which two diseases are caused by the pests and one due to pathogens (infectious organisms) and environmental conditions and also show the area damaged by a disease in leaves. Namely Red Rust, Helopeltis and Red spider mite respectively. In this

He Huang, Qi Yang, Yiling Xu, Zhu Li

Time-varying meshes, characterized by dynamic connectivity and varying vertex counts, hold significant promise for applications such as augmented reality. However, their practical utilization remains challenging due to the substantial data volume required for high-fidelity representation. While various compression methods attempt to leverage temporal redunda

Tien-Tien Yeh, Evan Wilson, Mikael Fogelström, Alexander Balatsky

We report a light-driven non-equilibrium vortex Berezinskii-Kosterlitz-Thouless (BKT) transition in a superconductor. We use a time-dependent Ginzburg-Landau model to demonstrate vortex-antivortex deconfinement via light induced fields. The transformation occurs independently of thermal fluctuations and is viewed as a quantum phase transition. The resulting

Weiyi Wang, Mahendra Piraveenan

This study examines how contact network topology influences the effectiveness of vaccination programs in the context of human papillomavirus (HPV) transmission. Using the SeCoNet sexual contact network growth model, we evaluate age based, ring based, and several centrality based vaccination strategies across the overall, male, and female cohorts, focusing on

Sofiane Ennadir, Oleg Smirnov, Yassine Abbahaddou, Lele Cao

Graph Neural Networks (GNNs) have achieved strong performance across a range of graph representation learning tasks, yet their adversarial robustness in graph classification remains underexplored compared to node classification. While most existing defenses focus on the message-passing component, this work investigates the overlooked role of pooling operatio

Hojun Choi, Eunho Heo, Dongjin Lee

Dimensionally decomposed generalized polynomial chaos expansion (DD-GPCE) efficiently performs forward uncertainty quantification (UQ) in complex engineering systems with high-dimensional random inputs of arbitrary distributions. However, constructing the measure-consistent orthonormal polynomial bases in DD-GPCE requires prior knowledge of input distributio

Eranga Bandara, Ross Gore, Xueping Liang, Sachini Rajapakse

The emergence of Agentic AI is fundamentally transforming how software is designed, developed, and maintained. Traditional software development methodologies such as Agile, Kanban, ShapeUp, etc, were originally designed for human-centric teams and are increasingly inadequate in environments where autonomous AI agents contribute to planning, coding, testing,

Divya Jyoti Bajpai, Manjesh Kumar Hanawal

Vision-language Models (VLMs) have made significant strides in visual understanding and query response generation, but often face challenges of high computational cost and inference latency due to autoregressive decoding. In this work, we introduce an imitation-learning-based Self-Speculative Decoding (SSD) framework, named FastVLM, to address these limitati

Soumya Halder, Koushik Gayen, Nagendra S. Kamath, Suman Kalyan Pal

Quasi-two-dimensional (quasi-2D) metal halide perovskites exhibit rich phase heterogeneity that profoundly influences light-matter interactions and charge transport. However, the fundamental mechanisms governing carrier transfer across distinct phases remain poorly understood. Here, we demonstrate effective electron-hole separation in fluorinated multilayere

Jia-Xue Niu, Yan-Nan Zhao, Rui Guo, Jian-Wen Zhang

In this work, we will use inverse scattering transform to study the semi-discrete Gardner equation under two types of non-vanishing boundary conditions, and investigate two interesting nonlinear waves in the presence of discrete spectrum, namely heteropolar solitons and kinks. When $u_n\rightarrow -\frac{a}{2b}$ as $n\rightarrow \pm \infty$, this is a symmet

Tenyo Takahashi

We generalize the theory of stable canonical rules by adopting definable filtration, a generalization of the method of filtration. We show that for a modal rule system or a modal logic that admits definable filtration, each extension is axiomatizable by stable canonical rules. Moreover, we provide an algebraic presentation of Gabbay's filtration and generali

Yuval Bar Ilan, Boaz Rafaely, Vladimir Tourbabin

Speech enhancement is a fundamental challenge in signal processing, particularly when robustness is required across diverse acoustic conditions and microphone setups. Deep learning methods have been successful for speech enhancement, but often assume fixed array geometries, limiting their use in mobile, embedded, and wearable devices. Existing array-agnostic

Paul Alexander Bilokon

We construct a sieve that enumerates rational ``imbalances'' of the form $(p-q)/(p+q)$ for integers $p\ge2$ and $1\le q<p$, ordered lexicographically by $(p,q)$. Each imbalance is reduced to lowest terms, and we record the sequence of distinct denominators as they first appear. We show that every positive integer occurs exactly once as such a denominator, an

Mateo Bouchet, Alejandro Tejedor, Xiangrong Wang, Yamir Moreno

We study diffusion on multiplex networks with directed interlayer couplings. We demonstrate both numerically and analytically that even with undirected layers, interlayer directionality alone reproduces superdiffusion and the prime regime. We further reveal a new phenomenon, the directionality-induced jamming, whereby directed interlayer links hinder diffusi

Marcos M. Cueli, Joaquín González-Nuevo, Laura Bonavera, Andrea Lapi

Weak lensing magnification probes the correlation between galaxies and the underlying matter field in a similar fashion to galaxy-galaxy lensing shear. Although it has long been sidelined in favor of the latter on the grounds of a poorer performance in terms of statistical significance, the provision of a large sample of high-redshift submillimeter galaxies

Jose Luis Ponton, Sheldon Andrews, Carlos Andujar, Nuria Pelechano

Interactive applications demand believable characters that respond naturally to dynamic environments. Traditional character animation techniques often struggle to handle arbitrary situations, leading to a growing trend of dynamically selecting motion-captured animations based on predefined features. While Motion Matching has proven effective for locomotion b

Marco De Santis, Lisa Alazraki

This paper presents our submission to the MRL 2025 Shared Task on Multilingual Physical Reasoning Datasets. The objective of the shared task is to create manually-annotated evaluation data in the physical commonsense reasoning domain, for languages other than English, following a format similar to PIQA. Our contribution, FormaMentis, is a novel benchmark for

Adinath Dukre, Ankan Deria, Yutong Xie, Imran Razzak

Atypical mitotic figures are important biomarkers of tumor aggressiveness in histopathology, yet reliable recognition remains challenging due to severe class imbalance and variability across imaging domains. We present a DenseNet-121-based framework tailored for atypical mitosis classification in the MIDOG 2025 (Track 2) setting. Our method integrates stain-

Ambalika Guha, Sajal Saha, Debanjan Ballav, Soumi Mitra

Preserving linguistic diversity is necessary as every language offers a distinct perspective on the world. There have been numerous global initiatives to preserve endangered languages through documentation. This paper is a part of a project which aims to develop a trilingual (Toto-Bangla-English) language learning application to digitally archive and promote

Md. Mehedi Hasan, Sk Tanzir Mehedi, Ziaur Rahman, Rafid Mostafiz

This paper presents a real-time modular defense system named Sentra-Guard. The system detects and mitigates jailbreak and prompt injection attacks targeting large language models (LLMs). The framework uses a hybrid architecture with FAISS-indexed SBERT embedding representations that capture the semantic meaning of prompts, combined with fine-tuned transforme

Mohd Faisal Khan, Mukul Lokhande, Santosh Kumar Vishvakarma

Edge-AI applications still face considerable challenges in enhancing computational efficiency in resource-constrained environments. This work presents RAMAN, a resource-efficient and approximate posit(8,2)-based Multiply-Accumulate (MAC) architecture designed to improve hardware efficiency within bandwidth limitations. The proposed REAP (Resource-Efficient A

Adhyayan Veer Singh, Aaron Shen, Brian Law, Ahmed Ismail

Correctness alone is insufficient: LLM-generated programs frequently satisfy unit tests while violating contest time or memory budgets. We present SwiftSolve, a complexity-aware multi-agent system for competitive programming that couples algorithmic planning with empirical profiling and complexity-guided repair. We frame competitive programming as a software

Raoul Serao, Gianpaolo Torre, Antonio Capolupo

This review explores the interplay between quantum information theory and high-energy physics, emphasizing how decoherence effects and unconventional neutrino oscillation patterns may unveil fundamental properties such as the Dirac or Majorana nature of neutrinos and potential CPT violation. It further discusses the use of entanglement measures as novel prob

Yuetong Luo

The signature of a closed manifold is an important geometric topology. Let $M$ be a closed manifold and $N$ be a codimension 2 submanifold of it. Given certain homotopy conditions, Higson, Xie and Schick proved an invariance theorem in codimension 2 for the $K$-theoretic signature. They asked for the $L$-theoretic counterpart of their result. In this note, w

Padmanabhan Jagannathan Prajesh, Kaliaperumal Ragunath, Miriam Gordon, Bruce Rathgeber

Accurate mapping of column-averaged CO2 (XCO2) over agricultural landscapes is essential for guiding emission mitigation strategies. We present a Spatiotemporal Vision Transformer with Wavelets (ST-ViWT) framework that reconstructs continuous, uncertainty-quantified XCO2 fields from OCO-2 across southern Canada, emphasizing poultry-intensive regions. The mod

Kangran Zhao, Yupeng Chen, Xiaoyu Zhang, Yize Chen

The misuse of advanced generative AI models has resulted in the widespread proliferation of falsified data, particularly forged human-centric audiovisual content, which poses substantial societal risks (e.g., financial fraud and social instability). In response to this growing threat, several works have preliminarily explored countermeasures. However, the la

Julia Bazinska, Max Mathys, Francesco Casucci, Mateo Rojas-Carulla

AI agents powered by large language models (LLMs) are being deployed at scale, yet we lack a systematic understanding of how the choice of backbone LLM affects agent security. The non-deterministic sequential nature of AI agents complicates security modeling, while the integration of traditional software with AI components entangles novel LLM vulnerabilities

Songhan Zhang, Yuanhao Lai, Pengfei Zheng, Boxi Yu

Multivariate time series (MTS) anomaly detection is essential for maintaining the reliability of industrial systems, yet real-world deployment is hindered by two critical challenges: training data contamination (noises and hidden anomalies) and inefficient model inference. Existing unsupervised methods assume clean training data, but contamination distorts l

Mackenzie Tapp, Sibi Chakravarthy Parivendan, Kashfia Sailunaz, Suresh Neethirajan

Pose estimation serves as a cornerstone of computer vision for understanding animal posture, behavior, and welfare. Yet, agricultural applications remain constrained by the scarcity of large, annotated datasets for livestock, especially dairy cattle. This study evaluates the potential and limitations of cross-species transfer learning by adapting ZebraPose -

Mariana F. Ramos, Costin Luchian, Michael Hentschel, Florian Honz

We deploy shortwave-QKD over short-reach in-house/datacom architectures and show that few-mode propagation and speckle-selective loss severely impact the QKD performance. We accomplish 12 kb/s secure-key generation in presence of 50 co-existing data channels.

Morteza Alikhani, Mohammadtaha Bagherifard, Erfan Zinvandi, Mehran Sarmadi

We introduced PerCoR (Persian Commonsense Reasoning), the first large-scale Persian benchmark for commonsense reasoning. PerCoR contains 106K multiple-choice sentence-completion problems drawn from more than forty news, cultural, and other web sources. We introduce a novel conjunction-based segmentation strategy to generate coherent sentence-completion pairs

Roham Koohestani, Agnia Sergeyuk, David Gros, Claudio Spiess

The introduction of large language models into integrated development environments (IDEs) is revolutionizing software engineering, yet it poses challenges to the usefulness and reliability of Artificial Intelligence-generated code. Post-hoc calibration of internal model confidences aims to align probabilities with an acceptability measure. Prior work suggest

Songhan Zhang, Aoyang Fang, Yifan Yang, Ruiyi Cheng

Cloud-native microservices enable rapid iteration and scalable deployment but also create complex, fast-evolving dependencies that challenge reliable diagnosis. Existing root cause analysis (RCA) approaches, even with multi-modal fusion of logs, traces, and metrics, remain limited in capturing dynamic behaviors and shifting service relationships. Three criti

Peter Danchev, Mina Doostalizadeh, Mehrdad Esfandiar, Omid Hasanzadeh

We investigate the so-called {\it $UJ^{\#}$ rings}, a new type of rings in which every unit can be written as $1+j$ with $j\in J^{\#}(R)$. These rings were defined and studied by Saini-Udar in Czechoslovak Math. J. (2025) under the name {\it $\sqrt{J}U$ rings}. (See \cite{SU}.) This class extends both the classes of UU and UJ rings, but also has its own spec

Md. Mehedi Hasan, Md. Abir Hossain, Farman Hossain Sayem, Bikash Kumar Paul

Accurate symptom-to-disease classification and clinically grounded treatment recommendations remain challenging, particularly in heterogeneous patient settings with high diagnostic risk. Existing large language model (LLM)-based systems often lack medical grounding and fail to quantify uncertainty, resulting in unsafe outputs. We propose CLIN-LLM, a safety-c

Harindu Jayarathne, Dileepa Marasinghe, Nandana Rajatheva, Matti Latva-aho

An end-to-end learning method for constellation shaping with a shaping-encoder assisted transceiver architecture is presented. The shaping encoder, which produces shaping bits with a higher probability of zeros, is used to produce an efficient symbol probability distribution. Both the probability distribution and the constellation geometry are jointly optimi

Yassh Ramchandani, Vijayashekhar S S, Jignesh S. Bhatt

In this article, we present SWAN: a three-stage, self-supervised wavelet neural network for joint estimation of endmembers and abundances from hyperspectral imagery. The contiguous and overlapping hyperspectral band images are first expanded to Biorthogonal wavelet basis space that provides sparse, distributed, and multi-scale representations. The idea is to

Liangzhe Han, Leilei Sun, Tongyu Zhu, Tao Tao

As a window for urban sensing, human mobility contains rich spatiotemporal information that reflects both residents' behavior preferences and the functions of urban areas. The analysis of human mobility has attracted the attention of many researchers. However, existing methods often address specific tasks from a particular perspective, leading to insufficien

Yuang Wang, Pengfei Jin, Siyeop Yoon, Matthew Tivnan

Reconstructing CT images from incomplete projection data remains challenging due to the ill-posed nature of the problem. Diffusion bridge models have recently shown promise in restoring clean images from their corresponding Filtered Back Projection (FBP) reconstructions, but incorporating data consistency into these models remains largely underexplored. Inco

Ghorbanali Haghighatdoost

This article provides a conceptual and historical review of the evolution of integrable Hamiltonian systems from the Moscow School of A. T. Fomenko to the emerging Azarbaijan School of Geometric Dynamical Systems founded by the author. Beginning with the topological classification of integrable systems through Liouville foliations, atoms, and molecular invar

Anand, Umberto Cappellazzo, Stavros Petridis, Maja Pantic

Large language models (LLMs) have recently advanced auditory speech recognition (ASR), visual speech recognition (VSR), and audio-visual speech recognition (AVSR). However, understanding of their internal dynamics under fine-tuning remains limited. In natural language processing, recent work has revealed attention sinks, tokens that attract disproportionatel

Mahyar Abbasian, Ramesh Jain

Building on decades of success in digital infrastructure and biomedical innovation, we propose the Personal Care Utility (PCU) - a cybernetic system for lifelong health guidance. PCU is conceived as a global, AI-powered utility that continuously orchestrates multimodal data, knowledge, and services to assist individuals and populations alike. Drawing on mult

Cook Hyun Kim, Hoyun Choi, Joonsung Jung, B. Kahng

Spin systems have emerged as powerful tools for understanding collective phenomena in complex systems. In this work, we investigate the Ashkin--Teller (AT) model on random scale-free networks using mean-field theory, which extends the traditional Ising framework by coupling two spin systems via both pairwise and four-spin interactions. We focus on the previo

Matthew Headrick, Sreeman Reddy Kasireddy, Andrew Rolph

We derive several new quantum bit thread prescriptions for holographic entanglement entropy, equivalent for static states to the quantum extremal surface formula. Our new prescriptions come in many varieties: vector field-based or based on measures over bulk curves, dependent or independent of the bulk UV regulator, loose and strict versions of constraints,

Huilin Yin, Zhaolin Yang, Linchuan Zhang, Gerhard Rigoll

The reliability of Simultaneous Localization and Mapping (SLAM) is severely constrained in environments where visual inputs suffer from noise and low illumination. Although recent 3D Gaussian Splatting (3DGS) based SLAM frameworks achieve high-fidelity mapping under clean conditions, they remain vulnerable to compounded degradations that degrade mapping and

Yasharth Yadav, Kelin Xia

Geometric data analysis and learning has emerged as a distinct and rapidly developing research area, increasingly recognized for its effectiveness across diverse applications. At the heart of this field lies curvature, a powerful and interpretable concept that captures intrinsic geometric structure and underpins numerous tasks, from community detection to ge

Peng Wang

Using a large observational sample from the Sloan Digital Sky Survey, we investigate the spatial alignment between galaxy pairs and their local cosmic filaments. Focusing on pairs with stellar masses and separations comparable to the Milky Way-Andromeda (MW-M31) system, we measure the angle between the pair connecting line and the orientation of the host fil

Cook Hyun Kim, Jinha Park, Young Jin Kim, Sangjoon Park

Globally coupled oscillator systems with inertia exhibit complex synchronization patterns, among which the emergence of a couple of secondary synchronized clusters (SCs) in addition to the primary cluster (PC) is especially distinctive. Although previous studies have predominantly focused on the collective properties of the PC, the dynamics of individual clu

Yoshiyasu Ito, Daisuke Kadoh, Yuki Sato

We investigate pinched geometries in a two-dimensional Lorentzian model of quantum Regge calculus (QRC) using the tensor renormalization group (TRG) method. A pinched geometry refers to a configuration with an infinitely long temporal extent, even when the total spacetime area is fixed. We examine several choices of integration measures and triangulations to

Bingqing Song, Jiaxiang Li, Rong Wang, Songtao Lu

Pre-trained large language models have demonstrated a strong ability to learn from context, known as in-context learning (ICL). Despite a surge of recent applications that leverage such capabilities, it is by no means clear, at least theoretically, how the ICL capabilities arise, and in particular, what is the precise role played by key factors such as pre-t

Dario Loi, Elena Maria Muià, Federico Siciliano, Giovanni Trappolini

We present AutoBench, a fully automated and self-sustaining framework for evaluating Large Language Models (LLMs) through reciprocal peer assessment. This paper provides a rigorous scientific validation of the AutoBench methodology, originally developed as an open-source project by eZecute S.R.L.. Unlike static benchmarks that suffer from test-set contaminat

S. Stalin, M. Lakshmanan

In this paper, we propose an alternative approach to generate a new class of beating vector solitons. Unlike earlier procedures that use dark-bright or bright-dark soliton solutions to generate beating solitons, the method described here utilizes non-degenerate vector soliton solutions of the Manakov system. It involves linear superposition of such soliton s

Yassir Lairgi, Ludovic Moncla, Khalid Benabdeslem, Rémy Cazabet

In today's rapidly expanding data landscape, knowledge extraction from unstructured text is vital for real-time analytics, temporal inference, and dynamic memory frameworks. However, traditional static knowledge graph (KG) construction often overlooks the dynamic and time-sensitive nature of real-world data, limiting adaptability to continuous changes. Moreo

Boyi Zheng, Yalin Zheng, Hrvoje Bogunović, Qing Liu

In this work, we propose PSScreen V2, a partially supervised self-training framework for multiple retinal disease screening. Unlike previous methods that rely on fully labelled or single-domain datasets, PSScreen V2 is designed to learn from multiple partially labelled datasets with different distributions, addressing both label absence and domain shift chal

Francesco Spinnato

Contemporary chess engines offer precise yet opaque evaluations, typically expressed as centipawn scores. While effective for decision-making, these outputs obscure the underlying contributions of individual pieces or patterns. In this paper, we explore adapting SHAP (SHapley Additive exPlanations) to the domain of chess analysis, aiming to attribute a chess

Wenming Tu, Guanrou Yang, Ruiqi Yan, Wenxi Chen

Spoken dialogue models currently lack the ability for fine-grained speech style control, a critical capability for human-like interaction that is often overlooked in favor of purely functional capabilities like reasoning and question answering. To address this limitation, we introduce UltraVoice, the first large-scale speech dialogue dataset engineered for m

Hexuan Zhi, Yanbo Zhang

For given graphs $G_1, \ldots, G_k$, let $\tilde{r}(G_1, \ldots, G_k)$ denote their online Ramsey number. In an influential paper on the online Ramsey numbers for paths and cycles, Cyman, Dzido, Lapinskas, and Lo (Electron. J. Combin., 2015) determined the exact values of $\tilde{r}(P_3, P_{\ell})$ and $\tilde{r}(P_3, C_{\ell})$. They also conjectured the ex

Noa Shoham, Ron Dorfman, Shalev Shaer, Kfir Y. Levy

Prediction-Powered Inference (PPI) is a recently proposed statistical inference technique for parameter estimation that leverages pseudo-labels on both labeled and unlabeled data to construct an unbiased, low-variance estimator. In this work, we extend its core idea to semi-supervised learning (SSL) for model training, introducing a novel unbiased gradient e

Thierry Daudé, Fabricio Macià, Cristóbal Meroño, François Nicoleau

The problem of characterizing sequences of real numbers that arise as spectra of Dirichlet-to-Neumann (DtN) maps for elliptic operators has attracted considerable attention over the past fifty years. In this article, we address this question in the simple setting of DtN maps associated with a rotation-invariant elliptic operator $\nabla \cdot (\gamma\nabla \

Giedrius Alkauskas

Fix an integer n>=1. Suppose that a simple polygon is the union of n triangles whose vertices along the common boundary are arranged cyclically. How many sides can such a union -- to be called regular -- have at most? This gives OEIS sequence A375986, a recent entry. It will be shown here that the sequence begins 3, 12, 22, 33, 45, 56, 67, 80, 91, and satisf

Ubaid Tantary

We quantify the transition between weak and strong coupling in thermal ${\cal N}=4$ supersymmetric Yang--Mills (SYM) theory in four space-time dimensions by constructing an \emph{admissible ensemble} of log-aware Pad\'e approximants that incorporate the weak- and strong-coupling expansions through $\mathcal O(\lambda^2)$ and $\mathcal O(\lambda^{-3/2})$ ($\l

Jian Sun, Kangdao Liu, Chi Zhang, Chuangquan Chen

Cross-view geo-localization (CVGL) plays a vital role in drone-based multimedia applications, enabling precise localization by matching drone-captured aerial images against geo-tagged satellite databases in GNSS-denied environments. However, existing methods rely on resource-intensive feature alignment and multi-branch architectures, incurring high inference

Chaitanya Shinde, Divya Garikapati

Generative Artificial Intelligence is emerging as a transformative force in the automotive industry, enabling novel applications across vehicle design, manufacturing, autonomous driving, predictive maintenance, and in vehicle user experience. This paper provides a comprehensive review of the current state of GenAI in automotive, highlighting enabling technol

Kaushal Kumar Maurya, Ekaterina Kochmar

The interdisciplinary research domain of Artificial Intelligence in Education (AIED) has a long history of developing Intelligent Tutoring Systems (ITSs) by integrating insights from technological advancements, educational theories, and cognitive psychology. The remarkable success of generative AI (GenAI) models has accelerated the development of large langu

Sayan Banik, Nikolai S. Kiselev, Ashis K. Nandy

Two-dimensional chiral magnets are known to host a variety of skyrmions, characterized by an integer topological charge. However, these systems typically favor uniform lattices as a thermodynamically stable phase composed of either skyrmions (Q = -1) or antiskyrmions (Q = 1). In isotropic chiral magnets, skyrmion-antiskyrmion coexistence is typically transie

Dhruv Sarkar, Abhishek Sinha

We propose an anytime online algorithm for the problem of learning a sequence of adversarial convex cost functions while approximately satisfying another sequence of adversarial online convex constraints. A sequential algorithm is called \emph{anytime} if it provides a non-trivial performance guarantee for any intermediate timestep $t$ without requiring prio