November 2025 arXiv papers — page 57

Simon Riche, Quan Situ

Let $G$ be a connected reductive algebraic group over an algebraically closed field of positive characteristic, $\mathfrak{g}$ be its Lie algebra, and $B$ be a Borel subgroup. We prove a formula for the dimensions of extension groups, in the principal block of the category of strongly $B$-equivariant $\mathfrak{g}$-modules (also called modular category $\mat

Khanh Gia Bui

Within the limited scope of this paper, we argue that artificial general intelligence cannot emerge from current neural network paradigms regardless of scale, nor is such an approach healthy for the field at present. Drawing on various notions, discussions, present-day developments and observations, current debates and critiques, experiments, and so on in be

Ravi Gupta, Shabista Haider

Smart home lighting systems consume 15-20% of residential energy but lack adaptive intelligence to optimize for user comfort and energy efficiency simultaneously. We present BitRL-Light, a novel framework combining 1-bit quantized Large Language Models (LLMs) with Deep Q-Network (DQN) reinforcement learning for real-time smart home lighting control on edge d

Haidong Kang, Ketong Qian, Yi Lu

Efforts to overcome catastrophic forgetting in Few-Shot Class-Incremental Learning (FSCIL) have primarily focused on developing more effective gradient-based optimization strategies. In contrast, little attention has been paid to the training cost explosion that inevitably arises as the number of novel classes increases, a consequence of relying on gradient

Ange-Clément Akazan, Issa Karambal, Jean Medard Ngnotchouye, Abebe Geletu Selassie. W

Physics-informed neural networks (PINNs) typically minimize average residuals, which can conceal large, localized errors. We propose Residual Risk-Aware Physics-Informed Neural Networks PINNs (RRaPINNs), a single-network framework that optimizes tail-focused objectives using Conditional Value-at-Risk (CVaR), we also introduced a Mean-Excess (ME) surrogate pe

Abishek Karthik, Sreya Mynampati, Pandiyaraju V

Solar energy is one of the most abundant and tapped sources of renewable energies with enormous future potential. Solar panel output can vary widely with factors like intensity, temperature, dirt, debris and so on affecting it. We have implemented a model on detecting dust and fault on solar panels. These two applications are centralized as a single-platform

He Huang, Yujun Guo, Wei He

Deep unfolding networks (DUNs) have achieved remarkable success and become the mainstream paradigm for spectral compressive imaging (SCI) reconstruction. Existing DUNs are derived from full-HSI imaging models, where each stage operates directly on the high-dimensional HSI, refining the entire data cube based on the single 2D coded measurement. However, this

Mihály A. Csirik, Andre Laestadius, Mathias Oster

Density functional theory is one of the most efficient and widely used computational methods of quantum mechanics, especially in fields such as solid state physics and quantum chemistry. From the theoretical perspecive, its central object is the universal density functional which contains all intrinsic information about the quantum system in question. Once t

Ashkan Javaherian

We present a MATLAB package for reconstructing sound-speed images from transmission ultrasound data. The package is based on two-point ray tracing and implements two complementary inversion strategies for image reconstruction. The first is a time-of-flight (ToF) method that produces low-resolution, low-contrast images with minimal artefacts. The second is a

Hong-Hanh Nguyen-Le, Van-Tuan Tran, Dinh-Thuc Nguyen, Nhien-An Le-Khac

The rapid advancement of generators (e.g., StyleGAN, Midjourney, DALL-E) has produced highly realistic synthetic images, posing significant challenges to digital media authenticity. These generators are typically based on a few core architectural families, primarily Generative Adversarial Networks (GANs) and Diffusion Models (DMs). A critical vulnerability i

Ziyu Meng, Tengyu Liu, Le Ma, Yingying Wu

Bipedal locomotion makes humanoid robots inherently prone to falls, causing catastrophic damage to the expensive sensors, actuators, and structural components of full-scale robots. To address this critical barrier to real-world deployment, we present \method, a framework that learns to predict imminent, unavoidable falls and execute protective maneuvers to m

Aliénor Rivière, David Fabre, Jacques Magnaudet, François Gallaire

Taking advantage of the recently developed L-ALE framework [Sierra-Ausin \textit{et al.}, Phys. Rev. Fluids {\bf{7}}, 113603 (2022)], we characterize the linear dynamics of an incompressible gas bubble immersed in a biaxial straining flow. We show that the system undergoes a saddle-node bifurcation with strongly different equilibrium shapes when varying the

Kai Jiang, Siqi Huang, Xiangyu Chen, Jiawei Shao

Multimodal large language models (MLLMs) deployed on devices must adapt to continuously changing visual scenarios such as variations in background and perspective, to effectively perform complex visual tasks. To investigate catastrophic forgetting under real-world scenario shifts, we construct a multimodal visual understanding dataset (MSVQA), covering four

Michael Adjei Osei, Sidney Shapiro

We study how to evaluate hybrid quantum programs as end-to-end workflows rather than as isolated devices or algorithms. Building on the Hybrid Quantum Program Evaluation Framework (HQPEF), we formalize a workflow-aware Quantum Readiness Level (QRL) score; define a normalized speedup under quality constraints for the Utility of Quantumness (UQ); and provide a

Wasilij Barsukow

Due to added numerical stabilization (diffusion), the stationary states of numerical methods for hyperbolic problems need not be consistent discretizations of those of the PDEs. A closely related phenomenon is the lack of consistency of common finite volume methods for the Euler equations in the limit of low Mach number. In this work, the stationary states o

Md Tasnin Tanvir, Soumitra Das, Sk Md Abidar Rahaman, Ali Shiri Sichani

The demand for edge AI in vision-language tasks requires models that achieve real-time performance on resource-constrained devices with limited power and memory. This paper proposes two adaptive compression techniques -- Sparse Temporal Token Fusion (STTF) and Adaptive Neural Compression (ANC) -- that integrate algorithmic innovations with hardware-aware opt

Aoi Wakuda

We provide some explicit algebraic criteria in terms of the Goldman bracket to decide whether two free homotopy classes of loops on an oriented surface admit disjoint representatives. We extend Kabiraj's method using the hyperbolic geometry of surfaces to prove these criteria. As an application, we show that the center of the Goldman Lie algebra of a pair of

Maciej Maliborski, Tobias C. Sutter

We study null geodesics that connect the two asymptotically flat regions of the maximally extended Kerr spacetime. These vortical geodesics traverse both horizons and pass through the ring singularity, linking the positive-$r$ exterior to the negative-$r$ asymptotic side. Using impact parameters, we identify a closed subset of parameter space, the inner thro

Chuqi Cao, Ling-Bing He, Yuanjie Lei, Qinghua Xiao

The physical essence of the non-relativistic limit, from the relativistic Vlasov-Maxwell-Landau system to the Vlasov-Poisson-Landau system, lies in the transition from finite-speed electromagnetic waves to instantaneous Coulomb interactions, and from relativistic to Newtonian particle dynamics. We rigorously justify this limit (mathematically corresponding t

Tyler Shoemaker

This position paper argues that literary scholars must engage with large language model (LLM) interpretability research. While doing so will involve ideological struggle, if not out-right complicity, the necessity of this engagement is clear: the abiding instrumentality of current approaches to interpretability cannot be the only standard by which we measure

Yue Li, Ifteher Alom, Wenhai Sun, Yang Xiao

Opening up data produced by the Internet of Things (IoT) and mobile devices for public utilization can maximize their economic value. Challenges remain in the trustworthiness of the data sources and the security of the trading process, particularly when there is no trust between the data providers and consumers. In this paper, we propose DEXO, a decentralize

Kevin Heng, Billy Edwards, Nicolas B. Cowan

Observations of transiting hot Jupiters have revealed a mismatch between the values of the Bond versus geometric albedos. In the planetary science literature, the ratio of these quantities is known as the phase integral. It has been extensively measured for the Solar System planets and shown to generally be non-unity in value. We use existing Cassini data of

Yi Zhang, Chao Zhang, Zijian Li, Tianxiang Xu

Large language models (LLMs) exhibit exceptional performance but pose substantial privacy risks due to training data memorization, particularly within healthcare contexts involving imperfect or privacy-sensitive patient information. We present a hierarchical dual-strategy framework for selective knowledge unlearning that precisely removes specialized knowled

Tomoyuki Yamakami

A computational model of adiabatic evolutionary quantum system (or AEQS, pronounced "eeh-ks") was introduced in [Yamakami,2022] as a sort of quantum annealing and its underlying input-driven Hamiltonians are generated quantum-algorithmically by various forms of quantum automata families (including 1qqaf's). We study an efficient way to accomplish certain mac

E. A. Ramirez Trino, M. A. Seifi MirJafarlou, M. A. Rajabpour

We develop a unified framework for Berezin integrals over Grassmann variables that establishes master identities for exponential quadratic fermionic forms and linear fermionic forms coupled to both bosonic and fermionic sources. The construction is rigorous for both real and complex fermions in arbitrary dimensions and remains well-defined even when the unde

Ziming Bian, Seung-Yeal Ha, Tommaso Ruggeri, Qinghua Xiao

Collective motion and self-organization of interacting particles, such as flocking and swarming, can be viewed as nonequilibrium analogues of collective dynamics in gases. Motivated by the analogy between gas mixtures and Cucker--Smale models, we introduce a polyatomic classical model and its relativistic counterpart based on the Synge energy, and analyze th

José Pombal, Maya D'Eon, Nuno M. Guerreiro, Pedro Henrique Martins

Demand for mental health support through AI chatbots is surging, though current systems present several limitations, like sycophancy or overvalidation, and reinforcement of maladaptive beliefs. A core obstacle to the creation of better systems is the scarcity of benchmarks that capture the complexity of real therapeutic interactions. Most existing benchmarks

Shenglan Yuan

L\'evy noise influences diverse non-equilibrium systems across scales, including quantum devices, active biological matter, and financial markets. While such noise is pervasive, its overall impact on activated transitions between metastable states remains unclear, despite prior studies of specific noise forms and scaling limits. In this work, we introduce a

Sai Puppala, Ismail Hossain, Md Jahangir Alam, Sajedul Talukder

Our study presents a multifaceted approach to enhancing user interaction and content relevance in social media platforms through a federated learning framework. We introduce personalized LLM Federated Learning and Context-based Social Media models. In our framework, multiple client entities receive a foundational GPT model, which is fine-tuned using locally

Samuel Ackerman, Wesam Ibraheem, Orna Raz, Marcel Zalmanovici

Systems incorporating large language models (LLMs) as a component are known to be sensitive (i.e., non-robust) to minor input variations that do not change the meaning of the input; such sensitivity may reduce the system's usefulness. Here, we present a framework to evaluate robustness of systems using COBOL code as input; our application is translation betw

Chunyu Qiang, Kang Yin, Xiaopeng Wang, Yuzhe Liang

Text-to-speech (TTS) and text-to-music (TTM) models face significant limitations in instruction-based control. TTS systems usually depend on reference audio for timbre, offer only limited text-level attribute control, and rarely support dialogue generation. TTM systems are constrained by input conditioning requirements that depend on expert knowledge annotat

Jasan Zughaibi, Denis von Arx, Maurus Derungs, Florian Heemeyer

Electromagnetic navigation systems (eMNS) enable a number of magnetically guided surgical procedures. A challenge in magnetically manipulating surgical tools is that the effective workspace of an eMNS is often severely constrained by power and thermal limits. We show that system-level control design significantly expands this workspace by reducing the curren

Daulet Berkimbayev

A straightforward and fully analytic approach is introduced to examine how polytropic fluids influence arbitrary gravitational sources in static, spherically symmetric spacetimes. As a concrete application, we explore the internal mechanism of energy transfer between gravitational sources embedded within a self-gravitating system.

Weiwei Chen, Huaxuan Xiao, Jiefeng Zhang, Xianjin Xia

LoRa has become a cornerstone for city-wide IoT applications due to its long-range, low-power communication. It achieves extended transmission by spreading symbols over multiple samples, with redundancy controlled by the Spreading Factor (SF), and further error resilience provided by Forward Error Correction (FEC). However, practical limits on SF and the sep

Benjamin A. Storer, Mehrnoush Kharghani, Alistair Adcroft, Hussein Aluie

Treatment of fields near domain boundaries is a long-standing problem in signal processing that has come into renewed focus following recent efforts in convolution-based multiscale coarse-graining and in machine-learned parameterizations due to ocean boundary artifacts. Here, we propose a general method for extending fields beyond the domain boundaries by so

Sergio Marin, Nhu Nguyen, Max, Zheng

We present a data-driven pipeline developed in collaboration with the Power Packs Project, a nonprofit addressing food insecurity in local communities. The system integrates data extraction from PDFs, large language models for ingredient standardization, and binary integer programming to generate a 15-week recipe schedule that minimizes projected wholesale c

Sara Venturini, Satyaki Sikdar, Martina Mazzarello, Francesco Rinaldi

The COVID-19 pandemic shifted academic collaboration from in-person to remote interactions. This study explores, for the first time, the effects on scientific collaborations and impact of such a shift, comparing research output before, during, and after the pandemic. Using large-scale bibliometric data, we track the evolution of collaboration networks and th

Zhijia Zhang, Mohsen Moazzami Gudarzi, Jiatong Mao, Ziwei Wang

Two-dimensional (2D) nanochannels have emerged as a pivotal platform for exploring nanoscale hydrodynamics and electrokinetics. Conventional fabrication methods to make nanochannels often introduce polymer contamination and require lengthy processing, limiting device performance and scalability. Here we introduce the Mask & Stack method, employing silicon ni

Jun-Da Pan, Yichao Li, Guo-Hong Du, Tian-Nuo Li

Understanding dark energy requires precision measurements of the expansion history of the universe and the growth of large-scale structure. The 21 cm intensity mapping (21 cm IM) technique enables rapid large-area surveys that can deliver these measurements. China is constructing three hundred-meter-class single-dish radio telescopes, including the QiTai 110

Jing Rou Puah, Sasa Arsovski

Accurate estimation of stellar parameters -- stellar age, lifetime, and evolutionary stage -- remains a fundamental challenge in astrophysics. We introduce a hybrid deep learning architecture combining multimodal spectroscopic and photometric data from SDSS DR17. The model comprises a Multi-Layer Perceptron for numerical features and a CNN with a Vision Tran

Tri Phu Vu

This paper studies choice situations in which a decision maker can choose multiple alternatives. Given a menu of available options, the decision maker selects a subset of the menu with certain probabilities. We employ an axiomatic approach to characterize various parametric models in the literature. Our results elucidate the implications of the functional fo

Bowen Zhao, Huanlai Xing, Zhiwen Xiao, Jincheng Peng

The attention mechanism has demonstrated remarkable potential in sequence modeling, exemplified by its successful application in natural language processing with models such as Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-trained Transformer (GPT). Despite these advancements, its utilization in time series forecasting (TS

M. E. Putman, D. A. Kim, S. E. Clark, L. Li

We present a catalog of 3D Galactic HI filaments over 1/3 of the sky using Galactic Arecibo L-band Feed Array HI (GALFA-HI) data. The 3D filaments are defined to be linear HI features that are continuous in position-position-velocity (PPV) and are found with fil3d, an algorithm that expands on the 2D FilFinder. The catalog contains 3333 HI filaments between

Winfried van den Dool, Maksim Zhdanov, Yuki M. Asano, Max Welling

Physical systems commonly exhibit spatially varying complexity, presenting a significant challenge for neural PDE solvers. While Graph Neural Networks can handle the irregular meshes required for complex geometries and boundary conditions, they still apply uniform computational effort across all nodes regardless of the underlying physics complexity. This lea

Juan Romero, Qiang Fu, Matteo Ravasi, Wolfgang Heidrich

Hyperspectral image reconstruction from a compressed measurement is a highly ill-posed inverse problem. Current data-driven methods suffer from hallucination due to the lack of spectral diversity in existing hyperspectral image datasets, particularly when they are evaluated for the metamerism phenomenon. In this work, we formulate hyperspectral image (HSI) r

Yves Tourigny

We consider the continuum limit of some products of random matrices in $\text{SL}(d,{\mathbb R})$ that arise as discretisations of incompressible renewing flows -- that is, of flows corresponding to a divergence-free velocity field that takes independent, identically-distributed values in successive time intervals of duration proportional to $\tau$. The stat

Heeseung Yun, Joonil Na, Jaeyeon Kim, Calvin Murdock

People continuously perceive and interact with their surroundings based on underlying intentions that drive their exploration and behaviors. While research in egocentric user and scene understanding has focused primarily on motion and contact-based interaction, forecasting human visual perception itself remains less explored despite its fundamental role in g

G. S. Bisnovatyi-Kogan

An approximate analytic one-dimensional model is constructed, for the accretion disk boundary layer surrounding a neutron star whose low magnetic field does not affect the process of accretion. A high luminosity model is considered, with radiation pressure dominant in the interior part of the disk.

Xiaoqing Wang, Keman Huang, Bin Liang, Hongyu Li

The rapid advancement of Large Language Model (LLM)-driven multi-agent systems has significantly streamlined software developing tasks, enabling users with little technical expertise to develop executable applications. While these systems democratize software creation through natural language requirements, they introduce significant security risks that remai

Metin Gürses, Aslı Pekcan

In this paper we study Hirota bilinear forms of the type $P(D) \{f\cdot f\}=0$. We prove that for $P(D)=D_x^mD_y^rD_t^n$ the equations have three-soliton solutions if only if two of nonzero $m,n,p$ are odd and the other one even. We explicitly derive the nonlinear partial differential equations corresponding to this form for $m+n+p=4$ and $m+n+p=6$. We show

Jack S. Calcut, Jamie Phillips-Freedman

We prove that each prime knot union an essential arc on a minimal genus Seifert surface is a prime theta-curve.

Jiayi Guo, Zijun Gao

We study the problem of selecting the best heterogeneous treatment effect (HTE) estimator from a collection of candidates in settings where the treatment effect is fundamentally unobserved. We cast estimator selection as a multiple testing problem and introduce a ground-truth-free procedure based on a cross-fitted, exponentially weighted test statistic. A ke

Bowei Pu, Chuanbin Liu, Yifan Ge, Peicheng Zhou

Video Large Language Models improve reasoning over complex videos by generating intermediate reasoning text. However, reliable reasoning depends on accurate video perception. In existing approaches, perception evidence is intertwined with reasoning text, making it difficult to directly supervise the perception process. We argue that reliable supervision requ

Shuo Wei, Haoyu Liu, Xin Sun, Youjin Deng

Criticality is traditionally regarded as an unstable, fine-tuned fixed point of the renormalization group. We introduce an iterative bicolored percolation process in two dimensions and show that it can both preserve criticality and transform fractal dimensions. Starting from critical configurations, such as the O$(n)$ loop and fuzzy Potts models, successive

Longyu Wu, Ji Shu

In this paper, we study the limiting behavior for stochastic differential equations driven by non-Gaussian alpha-stable Levy noise as alpha approaches 2. We first prove the convergence of solutions for system driven by alpha-stable Levy noise to those of the system driven by Brownian motion. Then we construct the C^1 inertial manifolds for both systems and s

Anupam Gupta, Vera Traub

In the Steiner Forest problem, we are given a graph with edge lengths, and a collection of demand pairs; the goal is to find a subgraph of least total length such that each demand pair is connected in this subgraph. For over twenty years, the best approximation ratio known for the problem was a $2$-approximation due to Agrawal, Klein, and Ravi (STOC 1991), d

Rui Liu, Xiangqi Qiang, Chengjun Hou

We study diagonal bimodules of \'{e}tale groupoid $C^*$-algebras over their canonical diagonal subalgebras, and establish necessary and sufficient conditions for such a bimodule to be spectral-that is, determined by its spectrum. For a class of $\Gamma$-graded \'{e}tale groupoids, we prove that the spectrality of diagonal bimodules is equivalent to their inv

Chrysafis Hartonas

We present a general relational semantics framework which, by varying the axiomatization and components of the relational structures, provides a uniform semantics for sentential logics, classical and non-classical alike. The approach we take rests on a generalization of the J\'{o}nsson-Tarski representation (and duality) for Boolean algebras with operators t

Duncan Stothers, Ben Stothers, Emily Schaeffer, Kishore Mulpuri

We study an ultrasound-first, radiation-preserving policy for developmental dysplasia of the hip (DDH) that requests a radiograph only when needed. We (i) pretrain modality-specific encoders (ResNet-18) with SimSiam on a large unlabelled registry (37186 ultrasound; 19546 radiographs), (ii) freeze the backbones and fit small, measurement-faithful heads on DDH

Yanbo Yin, Dingzhu Wen, Changsheng You, XiaoWen Cao

Space-Air-Ground Integrated Networks (SAGINs) are pivotal for enabling ubiquitous connectivity in 6G systems, yet they face significant challenges due to severe satellite-to-ground link impairments. Although Unmanned Aerial Vehicles (UAVs) can function as relay nodes to compensate for air-to-ground channel degradation, the satellite-to-UAV link remains a cri

Otome Obukohwo, Simon Jantač, Andrew Sowinski, Poupak Mehrani

The experiment of a single contact between two sub-centimeter high-speed particles is often difficult to execute, especially if the collision must be physically and electrically isolated, as is the case for triboelectrification studies. Apparatuses designed for this type of experiment fall short of providing high-speed isolated collisions with a high probabi

Diego Tuzi, Thomas Delamotte, Andreas Knopp

Direct-to-cell connectivity between satellites and common terrestrial handheld devices represents an essential feature of 6G. The industry is considering different type of constellations but using classical single satellite solutions based on phased array antennas. This article proposes to decompose a classical single satellite into a swarm of multiple small

Sami al-Asaad

We study the algebraic dynamics of endomorphisms of projective varieties. First, we characterize their iterated images, i.e. the intersection of the images of their iterates. Next, we explore the Stein factorizations of the iterates, proving some stability phenomena they exhibit. Finally, we study endomorphisms whose iterates lie in a finite union of connect

Loick Chambon, Paul Couairon, Eloi Zablocki, Alexandre Boulch

Vision Foundation Models (VFMs) extract spatially downsampled representations, posing challenges for pixel-level tasks. Existing upsampling approaches face a fundamental trade-off: classical filters are fast and broadly applicable but rely on fixed forms, while modern upsamplers achieve superior accuracy through learnable, VFM-specific forms at the cost of r

Susanne Pumpluen

Let $R=K[t;\sigma]$ be a skew polynomial ring, where $K$ is a cyclic Galois field extension of degree $n$ with Galois group generated by $\sigma$. We show that two irreducible similar skew polynomials $f,g\in R$ are similar if and only if they have the same bound. We prove that for two irreducible similar skew polynomials $f,g\in R$ the nonassociative Petit

Rui Xu, Dakuan Lu, Zicheng Zhao, Xiaoyu Tan

Spatial reasoning is a key capability in the field of artificial intelligence, especially crucial in areas such as robotics, computer vision, and natural language understanding. However, evaluating the ability of multimodal large language models(MLLMs) in complex spatial reasoning still faces challenges, particularly in scenarios requiring multi-step reasoni

Paniz Parastar, Giuseppe Caso, Jesus Alberto Omana Iglesias, Andra Lutu

Multi-access edge computing (MEC) is a promising solution for providing the computational resources and low latency required by vehicular services such as autonomous driving. It enables cars to offload computationally intensive tasks to nearby servers. Effective offloading involves determining when to offload tasks, selecting the appropriate MEC site, and ef

Shaoyu Liu, Jianing Li, Guanghui Zhao, Yunjian Zhang

Multimodal large language models (MLLMs) have made significant advancements in event-based vision, yet the comprehensive evaluation of their capabilities within a unified benchmark remains largely unexplored. In this work, we introduce EventBench, a benchmark that offers eight diverse task metrics together with a large-scale event stream dataset. EventBench

H. M. Schmid, J. Ma

Many circumstellar dust scattering regions have been detected and investigated with polarimetric imaging. However, the quantitative determination of the intrinsic polarization and of dust properties is difficult because of complex observational effects. This work investigates instrumental convolution and polarimetric calibration effect for high contrast imag

Mustapha Ouchen, Alex Prygarin

We analyze the experimental data of $\mathtt{p}\mathtt{-}\mathtt{p}$ collisions by the ATLAS and confront it with the AGK model developed by two of the authors, the Kharzeev-Levin~(KL) model and the simple exponential behavior for the Koba, Nielsen and Olesen~(KNO) scaling function. We show that all three models virtually coincide with all available experime

Vishesh Vishal Ahire, Yash Badrinarayan Amle, Akshada Nanasaheb Waditke, Ojas Nitin Ahire

The speed control security system is best suited for the task of slowing the speed of a vehicle during rash driving as the Driver is over speeding the circuit captures the images of the lanes witch decides the speed of the road the car is currently on this input is further provided to the ESP-32 micro Prosser module in the car switch compiles this data with

Christoph D. Dahl

Kaprekar's routine, i.e., sorting the digits of an integer in ascending and descending order and subtracting the two, defines a finite deterministic map on the state space of fixed-length digit strings. While its attractors (such as 495 for D = 3 and 6174 for D = 4) are classical, the global information-theoretic structure of the induced dynamics and its dep

Arpit Garg, Hemanth Saratchandran, Simon Lucey

Multimodal Large Language Models (MLLMs) increasingly need to forget specific knowledge such as unsafe or private information without requiring full retraining. However, existing unlearning methods often disrupt vision language alignment, causing models to reject both harmful and benign queries. We trace this failure to the projector network during unlearnin

Emanuele Longo, Josep Fontcuberta, Paolo Vavassori

Charge to spin (orbital) momentum conversion phenomena enclose great potential for advancing applications in spin/orbitronics. Although current-induced magnetic moment accumulation is crucial both for fundamental understanding and practical applications, direct quantifications are scarce. Optical polarization measurements, namely magneto-optical Kerr rotatio

Declan Nell, Milos Radonjic, Ivan Rungger, Liviu Chioncel

Two-terminal spintronic devices remain challenging to model under realistic operating conditions, where the interplay of complex electronic structures, correlation effects and bias-driven non-equilibrium dynamics may significantly impact charge and spin transport. Existing {\it ab initio} methods either capture bias-dependent transport but neglect dynamical

Lorenzo Rutayisire, Nicola Capodieci, Fabio Pellacini

Gaussian Splatting has emerged as a leading method for novel view synthesis, offering superior training efficiency and real-time inference compared to NeRF approaches, while still delivering high-quality reconstructions. Beyond view synthesis, this 3D representation has also been explored for editing tasks. Many existing methods leverage 2D diffusion models

S. K. Singh, Ahmed A. Zahia, Jia-Xin Peng, M. Y. Abd-Rabboud

We theoretically investigate a quantum battery architecture where two identical two-level atoms are charged by a cavity-magnomechanical system, which includes a microwave cavity, a magnon mode hosted in a YIG sphere, and phonon mode due to the deformation of the YIG sphere. The charging process relies on coherent energy exchange, where the atoms couple to th

Debapratim Banerjee, Debabrata Jana

In this paper, we show that the replica symmetry of the Gibbs measure of spherical spin systems is a property of the eigenvalue spacing at the edge of the interaction matrix. In particular, our interaction matrix has \textbf{two} large outlier eigenvalues with mutual distance $\frac{c}{n}$. The empirical measure of the rest of the eigenvalues is close to the

Xiangrui Zhang, Zeyu Chen, Haining Wang, Qiang Li

Large Language Models (LLMs) and their agent systems have recently demonstrated strong potential in automating code reasoning and vulnerability detection. However, when applied to large-scale firmware, their performance degrades due to the binary nature of firmware, complex dependency structures, and heterogeneous components. To address this challenge, this

Chi Zhang, Haibo Qiu, Qiming Zhang, Yufei Xu

Reinforcement Learning with Verifiable Rewards (RLVR) has significantly advanced the reasoning capabilities of Large Language Models (LLMs) and is now being applied to Vision-Language Models (VLMs). However, vanilla RLVR for VLMs verifies only the final textual output, critically neglecting the foundational step of visual perception. This oversight leads to

Tieguang Zi, Peng-Cheng Li, Bao-Min Gu, Fu-Wen Shu

We study the influence of a massive vector (Proca) field on the energy fluxes from extreme-mass-ratio inspirals (EMRIs) around a slowly rotating Kerr black hole. The secondary compact object, carrying a Proca hair, emits additional dipolar radiation that alters total energy flux relative to general relativity (GR). These modifications induce a secular drift

Yongkun Du, Pinxuan Chen, Xuye Ying, Zhineng Chen

The advent of Multimodal Large Language Models (MLLMs) has unlocked the potential for end-to-end document parsing and translation. However, prevailing benchmarks such as OmniDocBench and DITrans are dominated by pristine scanned or digital-born documents, and thus fail to adequately represent the intricate challenges of real-world capture conditions, such as

Yang Liu, Xiaolong Zhong, Ling Jiang

Large language models deliver strong reasoning and tool-use skills, yet their computational demands make them impractical for edge or cost-sensitive deployments. We present \textbf{Xmodel-2.5}, a 1.3-billion-parameter small language model designed as a \emph{drop-in agent core}. Training with maximal-update parameterization ($\mu$P) allows hyper-parameters t

Yanbing Zhang, Ruifeng Yuan, Lei Wu

Adjoint based shape optimization is a powerful technique in fluid-dynamics optimization, capable of identifying an optimal shape within only dozens of design iterations. However, when extended to rarefied gas flows, the computational cost becomes enormous because both the six dimensional primal and adjoint Boltzmann equations must be solved for each candidat

Serim Han, Jingru Zhang, Hoseung Song

Graph-based methods have shown particular strengths in change-point detection (CPD) tasks for high-dimensional nonparametric settings. However, existing CPD research has rarely addressed data with repeated measurements or local group structures. A common treatment is to average repeated measurements, which can result in the loss of important within-individua

Arno Fehm, Ariyan Javanpeykar

This is a survey of results on the Hilbert property of algebraic varieties, and variants of it.

Xinyan Le, Yao Zhu, Yulin Hu, Bin Han

Physical layer security (PLS) is a potential solution for secure and reliable transmissions in future Ultra-Reliable and Low-Latency Communications (URLLC). This work jointly optimizes redundant bits and blocklength allocation in practical round-trip transmission scenarios. To minimize the leakage-failure probability, a metric that jointly characterizes secu

Rouzbeh Aghaieebeiklavasani, Gholam Reza Rokni Lamouki

One of the most notable aspects of mathematical modeling is that it sheds light on the complexities arising from changes in parameters and their real-world implications, thus gaining better insight into the dynamics of economic, political, and security phenomena. Moreover, modifications to mathematical modeling will set the stage for embedding new features i

Shivansh Chhawri, Rahul Mahadik, Suparna Rooj

Large Language Models (LLMs) require substantial computational resources, making model compression essential for efficient deployment in constrained environments. Among the dominant compression techniques: knowledge distillation, structured pruning, and low-bit quantization, their individual effects are well studied, but their interactions and optimal sequen

Mansur Yerzhanuly

Pneumonia remains a leading global cause of mortality where timely diagnosis is critical. We introduce LungX, a novel hybrid architecture combining EfficientNet's multi-scale features, CBAM attention mechanisms, and Vision Transformer's global context modeling for enhanced pneumonia detection. Evaluated on 20,000 curated chest X-rays from RSNA and CheXpert,

Avishka Perera, Kumal Hewagamage, Saeedha Nazar, Kavishka Abeywardana

Image-to-point cross-modal learning has emerged to address the scarcity of large-scale 3D datasets in 3D representation learning. However, current methods that leverage 2D data often result in large, slow-to-train models, making them computationally expensive and difficult to deploy in resource-constrained environments. The architecture design of such models

Ziyuan Dong, Xiang Fan, Tengxun Zhong, Daowen Qiu

This work establishes a new probabilistic bound on the number of elements to generate finite nilpotent groups. Let $\varphi_k(G)$ denote the probability that $k$ random elements generate a finite nilpotent group $G$. For any $0 < \epsilon < 1$, we prove that $\varphi_k(G) \ge 1 - \epsilon$ if $k \ge \operatorname{rank}(G) + \lceil \log_2(2/\epsilon) \rceil$

B. Y. Yan, Chaofan Li, Hongjin Qian, Shuqi Lu

Memory is critical for AI agents, yet the widely-adopted static memory, aiming to create readily available memory in advance, is inevitably subject to severe information loss. To address this limitation, we propose a novel framework called \textbf{general agentic memory (GAM)}. GAM follows the principle of "\textbf{just-in time (JIT) compilation}" where it f

Mohammad Jafari Vayeghan, Niloufar Delfan, Mehdi Tale Masouleh, Mansour Parvaresh Rizi

Precise 3D segmentation of cerebral vasculature from T1-weighted contrast-enhanced (T1CE) MRI is crucial for safe neurosurgical planning. Manual delineation is time-consuming and prone to inter-observer variability, while current automated methods often trade accuracy for computational cost, limiting clinical use. We present NeuroVascU-Net, the first deep le

Georgios C. Chasparis

Reinforcement-based learning dynamics may exhibit several limitations when applied in a distributed setup. In (repeatedly-played) multi-player/action strategic-form games, and when each player applies an independent copy of the learning dynamics, convergence to (usually desirable) pure Nash equilibria cannot be guaranteed. Prior work has only focused on a sm

Yoshihiro Maruyama

We develop a theory of category-equivariant neural networks (CENNs) that unifies group/groupoid-equivariant networks, poset/lattice-equivariant networks, graph and sheaf neural networks. Equivariance is formulated as naturality in a topological category with Radon measures. Formulating linear and nonlinear layers in the categorical setup, we prove the equiva

Haonan Wang, Hanyu Zhou, Haoyue Liu, Luxin Yan

We investigate a challenging task of dynamic scene geometry estimation, which requires representing both spatial and temporal features. Typically, existing methods align the two features into a unified latent space to model scene geometry. However, this unified paradigm suffers from potential mismatched representation due to the heterogeneous nature between

Wei Yang, Yiran Zhu, Zilin Li, Xunjia Zhang

Coarse-to-fine path decision-making requires predicting a valid taxonomy path in which earlier decisions constrain later ones. However, existing benchmarks score each level independently, obscuring cross-level validity and consistency. To better align evaluation with this setting, we introduce a Joint Path Decision (JPD) protocol that requires predicting the

Dingli Yuan, Jingchen Peng, Jie Fan, Boxiang Ren

The wireless communication environment has the characteristic of strong dynamics. Conventional wireless networks operate based on the static rules with predefined algorithms, lacking the self-adaptation ability. The rapid development of artificial intelligence (AI) provides a possibility for wireless networks to become more intelligent and fully automated. A

Chris Kuchar

RFX (Random Forests X), where X stands for compression or quantization, presents a production-ready implementation of Breiman and Cutler's Random Forest classification methodology in Python. RFX v1.0 provides complete classification: out-of-bag error estimation, overall and local importance measures, proximity matrices with QLORA compression, case-wise analy