October 2025 arXiv papers — page 7

Mengshi Qi, Jiaxuan Peng, Jie Zhang, Juan Zhu

In the machine learning system, the hybrid model parallelism combining tensor parallelism (TP) and pipeline parallelism (PP) has become the dominant solution for distributed training of Large Language Models~(LLMs) and Multimodal LLMs (MLLMs). However, TP introduces significant collective communication overheads, while PP suffers from synchronization ineffic

Xin Tang, Youfang Han, Fangfei Gou, Wei Zhao

Vision-Language Models (VLMs) excel in diverse multimodal tasks. However, user requirements vary across scenarios, which can be categorized into fast response, high-quality output, and low energy consumption. Relying solely on large models deployed in the cloud for all queries often leads to high latency and energy cost, while small models deployed on edge d

Yehna Kim, Young-Eun Kim, Seong-Whan Lee

Vision-Language Models (VLMs) have demonstrated impressive capabilities in zero-shot action recognition by learning to associate video embeddings with class embeddings. However, a significant challenge arises when relying solely on action classes to provide semantic context, particularly due to the presence of multi-semantic words, which can introduce ambigu

Rajan Agarwal, Aarush Gupta

Instruction-tuned Large Language Models (LLMs) underperform on low resource, non-Latin scripts due to tokenizer fragmentation and weak cross-lingual coupling. We present LLINK (Latent Language Injection for Non-English Knowledge), a compute efficient language-as-modality method that conditions an instruction-tuned decoder without changing the tokenizer or re

Qiyan Deng, Changqian Zheng, Lianpeng Qiao, Yuping Wang

Dataset distillation condenses large datasets into synthetic subsets, achieving performance comparable to training on the full dataset while substantially reducing storage and computation costs. Most existing dataset distillation methods assume that all real instances contribute equally to the process. In practice, real-world datasets contain both informativ

Fatemah Ayatollah Zadeh Shirazi, Sajjad Moradi Chaleshtori

In the following text we show if $X$ is an Alexandroff space, then $f:X\to Y$ has closed graph if and only if it has constant closed value on each connected component of $X$. Moreover, if $X$ an Alexandroff space and $f:X\to Y$ has closed graph, then $f:X\to Y$ is continuous. As a matter of fact, the number of maps which have closed graph from Alexandroff sp

Bijia Liu, Ronghao Dang

The exceptional potential of large language models (LLMs) in handling text information has garnered significant attention in the field of financial trading. However, most existing trading agents operate under intraday, independent unit-based trading tasks, where decisions are made as isolated directional actions, and thus lack awareness of continuous positio

Xin Jiang, Haitang Yang

We explicitly show how to derive the $(D+1)$-dimensional Einstein equation from the entanglement entropy between codimension-one {\it disjoint} regions in $D$-dimensional conformal field theory.

Ronja Rönnback, Chris Emmery, Marie Šafář Postma, Filip Milde

Search engines are used and trusted by hundreds of millions of people every day. However, the algorithms used by search engines to index, filter, and rank web content are inherently biased, and will necessarily prefer some views and opinions at the expense of others. In this article, we examine how these algorithmic biases amplify and suppress polarizing con

Deok-Seon Kim, Seo-Hyun Lee, Kang Yin, Seong-Whan Lee

Brain-to-speech (BTS) systems represent a groundbreaking approach to human communication by enabling the direct transformation of neural activity into linguistic expressions. While recent non-invasive BTS studies have largely focused on decoding predefined words or sentences, achieving open-vocabulary neural communication comparable to natural human interact

Mohammad Tavakoli, Alireza Salemi, Carrie Ye, Mohamed Abdalla

Evaluating the abilities of large language models (LLMs) for tasks that require long-term memory and thus long-context reasoning, for example in conversational settings, is hampered by the existing benchmarks, which often lack narrative coherence, cover narrow domains, and only test simple recall-oriented tasks. This paper introduces a comprehensive solution

Alik Pramanick, Mayank Bansal, Utkarsh Srivastava, Suklav Ghosh

In recent times, deep neural networks (DNNs) have been successfully adopted for various applications. Despite their notable achievements, it has become evident that DNNs are vulnerable to sophisticated adversarial attacks, restricting their applications in security-critical systems. In this paper, we present two-phase training methods to tackle the attack: f

Ora Nova Fandina, Gal Amram, Eitan Farchi, Shmulik Froimovich

Application modernization in legacy languages such as COBOL, PL/I, and REXX faces an acute shortage of resources, both in expert availability and in high-quality human evaluation data. While Large Language Models as a Judge (LaaJ) offer a scalable alternative to expert review, their reliability must be validated before being trusted in high-stakes workflows.

Jiachen Zhao, Xiao Yan, Eric Lo

Approximate $k$-nearest neighbor search (A$k$-NNS) is a core operation in vector databases, underpinning applications such as retrieval-augmented generation (RAG) and image retrieval. In these scenarios, users often prefer diverse result sets to minimize redundancy and enhance information value. However, existing greedy-based diverse methods frequently yield

Miguel Lagos

A polycrystalline solid is modelled as an ensemble of random irregular polyhedra filling the entire space occupied by the solid body, leaving no voids or flaws between them. Adjacent grains can slide with a relative velocity proportional to the local shear stress resolved in the plane common to the two sliding grains, provided it exceeds a threshold. The loc

Jingrui Zhang, Yimeng Xu, Shujie Li, Feng Liang

Federated Learning (FL) enables collaborative model training across decentralized clients without sharing private data. However, FL suffers from biased global models due to non-IID and long-tail data distributions. We propose \textbf{FedSM}, a novel client-centric framework that mitigates this bias through semantics-guided feature mixup and lightweight class

Sintu Kumar, Pratibha Jaiswal, Km. Shivani Bhardwaj, Rajesh Kumar Rai

Ultra-high field intensities are essential for developing high-energy-density physics and compact plasma accelerators, but they are essentially constrained by the limitations of focusing distance and nonlinear efficiency. We present a theoretical model for extreme laser energy concentration in under-dense plasma that shows a highly effective, magnetically su

Chuxuan Hu, Maxwell Yang, James Weiland, Yeji Lim

Manually conducting real-world data analyses is labor-intensive and inefficient. Despite numerous attempts to automate data science workflows, none of the existing paradigms or systems fully demonstrate all three key capabilities required to support them effectively: (1) open-domain data collection, (2) structured data transformation, and (3) analytic reason

Zhidong Yang, Xiuhui Shi, Wei Ba, Zhigang Song

Whole slide image (WSI) analysis has emerged as an increasingly essential technique in computational pathology. Recent advances in the pathology foundation models (FMs) have demonstrated significant advantages in deriving meaningful patch-level or slide-level multi-scale features from WSIs. However, current pathology FMs have exhibited substantial heterogene

Tianli Liao, Ran Wang, Siqing Zhang, Lei Li

Eliminating geometric distortion in semantically important regions remains an intractable challenge in image retargeting. This paper presents Object-IR, a self-supervised architecture that reformulates image retargeting as a learning-based mesh warping optimization problem, where the mesh deformation is guided by object appearance consistency and geometric-p

Tianyang Chu, Xiaoying Dai, Jing Wu, Aihui Zhou

This paper concerns the mathematical and numerical analysis of the $L^2$ normalized gradient flow model for the Gross--Pitaevskii eigenvalue problem, which has been widely used to design the numerical schemes for the computation of the ground state of the Bose--Einstein condensate. We first provide the mathematical analysis for the model, including the well-

Jingnan Gao, Zhe Wang, Xianze Fang, Xingyu Ren

Recent advances in language and vision have demonstrated that scaling up model capacity consistently improves performance across diverse tasks. In 3D visual geometry reconstruction, large-scale training has likewise proven effective for learning versatile representations. However, further scaling of 3D models is challenging due to the complexity of geometric

Liyang He, Zhenya Huang, Cheng Yang, Rui Li

With the rapid growth of textual content on the Internet, efficient large-scale semantic text retrieval has garnered increasing attention from both academia and industry. Text hashing, which projects original texts into compact binary hash codes, is a crucial method for this task. By using binary codes, the semantic similarity computation for text pairs is s

H. W. Qu, H. T. Liu, Y. C. Li

The excitonic insulator (EI) is a more than 60-year-old theoretical proposal that yet remains elusive. It is a purely quantum phenomenon involving the spontaneous generation of excitons in quantum mechanics and the spontaneous condensation of excitons in quantum statistics. At this point, the excitons represent the ground state rather than the conventional e

Partha Goswami

We present a theoretical study of the anomalous Nernst and Hall conductance in the Kagome metal potassium tri vanadium pentantimonide, based on a system Hamiltonian incorporating nearest neighbour and complex next nearest neighbour hopping, Rashba spin orbit coupling, an exchange field induced by magnetic proximity, and a charge density wave potential. Our a

Ali Nour Eldin, Benjamin Dalmas, Walid Gaaloul

In recent years, there has been a growing interest in the verification of business process models. Despite their lack of formal characterization, these models are widely adopted in both industry and academia. To address this issue, formalizing the execution semantics of business process modeling languages is essential. Since data and process are two facets o

Imteaz Rahaman, Michael E. Liao, Ziqi Wang, Eugene Y. Kwon

Rutile germanium dioxide (r-GeO2) has recently emerged as a promising ultrawide-bandgap (UWBG) semiconductor owing to its wide bandgap (~4.4-5.1 eV), ambipolar doping potential, and high theoretical thermal conductivity. However, experimental data on the thermal conductivity of r-GeO2 epitaxial layers have not been reported, primarily due to challenges in ph

Shaohong Li, Yi Zheng

We perform the first joint analysis of galaxy clustering (GC) and the kinetic Sunyaev-Zel'dovich (kSZ) effect to simultaneously constrain cosmological and astrophysical parameters in this work, utilizing a combination of the Atacama Cosmology Telescope (ACT) Data Release 6 (DR6) map and the Constant Stellar Mass (CMASS) galaxy sample. As a complementary prob

Shoma Yagi, Jun Ichikawa, Genki Ichinose

In football, attacking teams attempt to break through the opponent's defensive line to create scoring opportunities. This action, known as a Line Break, is a critical indicator of offensive effectiveness and tactical performance, yet previous studies have mainly focused on shots or goal opportunities rather than on how teams break the defensive line. In this

Ke Song, Chengcheng Ling, Haiyi Wang

We study a singular stochastic equation driven by a regular noise of fractional Brownian type with Hurst index $H \in (1,\infty)\setminus\mathbb{Z}$ and drift coefficient $b \in \mathcal{C}^\alpha$, where $\alpha > 1 - \frac{1}{2H}$. The strong well-posedness of this equation was first established in [Ger23], a phenomenon referred to as regularization by reg

Mahmoud El Hussieni, Bahadır K. Güntürk, Hasan F. Ateş, Oğuz Hanoğlu

Accurate building instance segmentation and height classification are critical for urban planning, 3D city modeling, and infrastructure monitoring. This paper presents a detailed analysis of YOLOv11, the recent advancement in the YOLO series of deep learning models, focusing on its application to joint building extraction and discrete height classification f

Tonmoy Deka, Liton Majumdar, Tasneem Basra Khan, Swastik Dewan

Thermal emission spectra provide key insights into the atmospheric composition and especially the temperature structure of an exoplanet. With broader wavelength coverage, sensitivity and higher resolution, JWST has enabled robust constraints on these properties, including detections of photochemical products. This advances the need for retrieval frameworks c

Jaebyeong Jeon, Hyeonseo Jang, Jy-yong Sohn, Kibok Lee

Equivariant representation learning aims to capture variations induced by input transformations in the representation space, whereas invariant representation learning encodes semantic information by disregarding such transformations. Recent studies have shown that jointly learning both types of representations is often beneficial for downstream tasks, typica

Zubiao Xiao, Hongwei Jia

In this paper, we introduce the notions of neutralized packing pressures and neutralized measure-theoretic pressures on subsets for a finitely generated free semigroup action. Let $X$ be a compact metric space and $\mathcal{G}$ be a finite family of continuous self-maps on $X$. We consider the semigroup $G$ generated by $\mathcal{G}$ on $X$. We show that the

Yixiang Chen, Tianshi Zheng, Shijue Huang, Zhitao He

Test-time scaling without interpreter feedback is essential for real-world code generation scenarios where test cases are not readily available. While existing paradigms often rely on either greedy exploitation (i.e., iterative refinement) or stochastic exploration (i.e., relying on sample-based voting or reranking mechanisms), the balance between these two

Yow-Ming Robin Hu, Elena A. Ostrovskaya, Eliezer Estrecho

In this work, we analytically derive a semi-classical equation of motion describing the zitterbewegung effects arising in the dynamics of wavepackets in non-Hermitian systems. In Hermitian non-relativistic quantum systems, the zitterbewegung effects can arise due to the spin precession and spin-orbit coupling. Interestingly, the spin dynamics in non-Hermitia

Renjie Ji, Xue Wang, Chao Niu, Wen Zhang

Hyperspectral imaging (HSI) is a critical technique for fine-grained land-use and land-cover (LULC) mapping. However, the inherent heterogeneity of HSI data, particularly the variation in spectral channels across sensors, has long constrained the development of model generalization via transfer learning or joint training. Existing HSI foundation models show

Sung Kim

We establish an identification between the spaces of $\alpha$-fusion trees in non-semisimple topological quantum computation (NSS TQC) and a family of homological representations of the braid group known as the Lawrence representations specialized at roots of unity. Leveraging this connection, we provide a new proof of Ito's colored Alexander invariant formu

Meijie Zhao, Xiao Wen

In this paper, we introduce the notions of rescaled metric pressure and rescaled topological pressure for flows by considering three types of rescaled Bowen balls, which take the flow velocity and time reparametrization into account. This approach effectively eliminates the influence of singularities. It is demonstrated that defining both metric pressure and

Shun-Qing Shen

The parity anomaly for Dirac fermions in two spatial dimensions has shaped perspectives in quantum field theory and condensed matter physics. In condensed matter it has evolved as a mechanism for half-quantized Hall responses in systems described by massive Dirac fermions. Here we reexamine the issue on a lattice and show that the half-quantized Hall conduct

Callum Dewsnap, Pauline Barmby, Sarah C. Gallagher

The ability to disentangle the light of an AGN from its host galaxy is strongly dependent on the spatial resolution and depth of the imaging. As the capabilities of imaging systems improve with time, confirming that our standard techniques adequately model the increasingly complex structures unveiled is essential. With JWST providing unprecedented image qual

Ren Tasai, Guang Li, Ren Togo, Takahiro Ogawa

We propose a novel continual self-supervised learning (CSSL) framework for simultaneously learning diverse features from multi-window-obtained chest computed tomography (CT) images and ensuring data privacy. Achieving a robust and highly generalizable model in medical image diagnosis is challenging, mainly because of issues, such as the scarcity of large-sca

Toshio Yanagida, Keisuke Fujita, Mitsuhiro Iwaki

The remarkable progress of artificial intelligence (AI) has revealed the enormous energy demands of modern digital architectures, raising deep concerns about sustainability. In stark contrast, the human brain operates efficiently on only ~20 watts, and individual cells process gigabit-scale genetic information using energy on the order of trillionths of a wa

Tao Liu, Chongyu Wang, Rongjie Li, Yingchen Yu

While Multimodal Large Language Models (MLLMs) have advanced GUI navigation agents, current approaches face limitations in cross-domain generalization and effective history utilization. We present a reasoning-enhanced framework that systematically integrates structured reasoning, action prediction, and history summarization. The structured reasoning componen

Spencer Daugherty, Nicolle González, Bárbara Muniz, Pablo S. Ocal

We study the monoid algebra ${}_{n}\mathcal{T}_{m}$ of semistandard Young tableaux, which coincides with the Gelfand--Tsetlin semigroup ring $\mathcal{GT}_{n}$ when $m = n$. Among others, we show that this algebra is commutative, Noetherian, reduced, Koszul, and Cohen--Macaulay. We provide a complete classification of its maximal ideals and compute the topol

Jiaxin Zhang, Zehong Zhu, Junye Deng, Yunqin Li

Villages areas hold significant importance in the study of human-land relationships. However, with the advancement of urbanization, the gradual disappearance of spatial characteristics and the homogenization of landscapes have emerged as prominent issues. Existing studies primarily adopt a single-disciplinary perspective to analyze villages spatial morpholog

Hamed Najafi, Dongsheng Luo, Jason Liu

Explainable AI (XAI) is critical for building trust in complex machine learning models, yet mainstream attribution methods often provide an incomplete, static picture of a model's final state. By collapsing a feature's role into a single score, they are confounded by non-linearity and interactions. To address this, we introduce Feature-Function Curvature Ana

Kounianhua Du, Jianxing Liu, Kangning Zhang, Wenxiang Jiao

The rapid evolution of large language models (LLMs) has intensified the demand for effective personalization techniques that can adapt model behavior to individual user preferences. Despite the non-parametric methods utilizing the in-context learning ability of LLMs, recent parametric adaptation methods, including personalized parameter-efficient fine-tuning

L. Yıldız, D. Kaykı, M. F. Ciappina

Geometry can fundamentally govern the propagation of light, independent of material constraints. Here, we demonstrate that a fractal phase space, endowed with a non-Euclidean, scale-dependent geometry, can intrinsically induce resonance quantization, spatial confinement, and tunable damping without the need for material boundaries or external potentials. Emp

Arthur Charpentier, Qiheng Guo, Mike Ludkovski

We analyze loss development in NAIC Schedule P loss triangles using functional data analysis methods. Adopting the functional viewpoint, our dataset comprises 3300+ curves of incremental loss ratios (ILR) of workers' compensation lines over 24 accident years. Relying on functional data depth, we first study similarities and differences in development pattern

Young-Seok Kweon, Gi-Hwan Shin, Ji-Yong Kim, Bokyeong Ryu

Conscious state estimation is important in various medical settings, including sleep staging and anesthesia management, to ensure patient safety and optimize health outcomes. Traditional methods predominantly utilize electroencephalography (EEG), which faces challenges such as high sensitivity to noise and the requirement for controlled environments. In this

Zhipeng Zhu, Lok Ming Lui

We propose a novel method for parameterizations of triangle meshes by finding an optimal quasiconformal map that minimizes an energy consisting of a relative entropy term and a quasiconformal term. By prescribing a prior probability measure on a given surface and a reference probability measure on a parameter domain, the relative entropy evaluates the differ

Krishan Kumar, P. Danumjaya, Anil Kumar, Amiya K. Pani

This paper deals with the asymptotic behavior and FEM error analysis of a class of strongly damped wave equations using a semidiscrete finite element method in spatial directions combined with a finite difference scheme in the time variable. For the continuous problem under weakly and strongly damping parameters $\alpha$ and $\beta,$ respectively, a novel ap

Jaydeep Mandal, Hartmut Löwen, Prabal K. Maiti

Using particle-resolved molecular-dynamics simulations, we compute the phase diagram for soft repulsive spherocylinders confined on the surface of a sphere. While crystal (K), smectic (Sm), and isotropic (I) phases exhibit a stability region for any aspect ratio of the spherocylinders, a nematic phase emerges only beyond a critical aspect ratio lying between

Bikash K. Das, Camilo Granados, Marcelo F. Ciappina

Light is the fundamental medium through which we perceive the world around us. In the modern era, light can not only be used in its raw form but can also be used as a versatile tool. Generally, light fields carry energy and momentum (both linear and angular). Due to the transfer of linear momentum from light to matter, the radiation pressure is exerted, wher

Aswin Balasubramanian, Jacques Distler, Ron Donagi, Carlos Perez-Pardavila

The Coulomb branch geometry of a 4d $\mathcal{N}=2$ SCFT is encoded in the data of a complex integrable system. In class-S, this is the Hitchin System (of ADE type) on the punctured curves $C$ on which we compactified from 6d to 4d. As we vary the complex structure of $C$, these fit together to form a (nontrivial!) bundle of Hitchin systems over the moduli s

Krishna Kingkar Pathak

Proton tunneling across hydrogen bonds is a fundamental quantum effect with implications for spectroscopy, catalysis, and biomolecular stability. While state-of-the-art instanton and path-integral methods provide accurate multidimensional tunneling splittings, simplified one-dimensional models remain valuable as conceptual and benchmarking tools. Here we dev

Anselm Lohmann, Tomohiro Nakatani, Rintaro Ikeshita, Marc Delcroix

Guided Source Separation (GSS) is a popular front-end for distant automatic speech recognition (ASR) systems using spatially distributed microphones. When considering spatially distributed microphones, the choice of reference microphone may have a large influence on the quality of the output signal and the downstream ASR performance. In GSS-based speech enha

Ziyuan Gao

Traffic accidents represent a critical public health challenge, claiming over 1.35 million lives annually worldwide. Traditional accident prediction models treat road segments independently, failing to capture complex spatial relationships and temporal dependencies in urban transportation networks. This study develops MDAS-GNN, a Multi-Dimensional Attention-

Zixin Chen, Hongzhan Lin, Kaixin Li, Ziyang Luo

The proliferation of memes on social media necessitates the capabilities of multimodal Large Language Models (mLLMs) to effectively understand multimodal harmfulness. Existing evaluation approaches predominantly focus on mLLMs' detection accuracy for binary classification tasks, which often fail to reflect the in-depth interpretive nuance of harmfulness acro

Zheng Li, Peng Zhang, Shixiao Liang, Hang Zhou

Understanding how Advanced Driver-Assistance Systems (ADAS) interact with Traffic Control Devices (TCDs) is critical for assessing their influence on traffic operations, yet this interaction has received little focused empirical study. This paper presents a field dataset and behavioral analysis of Tesla's Traffic Light and Stop Sign Control (TLSSC), a mature

Caixin Kang, Yifei Huang, Liangyang Ouyang, Mingfang Zhang

As AI systems become increasingly integrated into human lives, endowing them with robust social intelligence has emerged as a critical frontier. A key aspect of this intelligence is discerning truth from deception, a ubiquitous element of human interaction that is conveyed through a complex interplay of verbal language and non-verbal visual cues. However, au

Vahid Salehi, Josef Vilsmeier, Shirui Wang

Today, products are no longer isolated artifacts, but nodes in networked systems. This means that traditional, linearly conceived life cycle models are reaching their limits: Interoperability across disciplines, variant and configuration management, traceability, and governance across organizational boundaries are becoming key factors. This collective contri

Meng Li

We study Hamiltonian diffeomorphisms on symplectic Euclidean spaces that are equal to non-degenerate linear maps at infinity. Under the assumption that there exists an isolated homologically nontrivial fixed point satisfying the twist condition, we prove the existence of infinitely many simple periodic points. More precisely, if such a diffeomorphism has onl

Md Selim Sarowar, Sungho Kim

The primary challenge in computer vision is precisely calculating the pose of 6D objects, however many current approaches are still fragile and have trouble generalizing from synthetic data to real-world situations with fluctuating lighting, textureless objects, and significant occlusions. To address these limitations, VLM6D, a novel dual-stream architecture

Mengbo Wang, Shourya Verma, Aditya Malusare, Luopin Wang

Spatial transcriptomics (ST) technologies can be used to align transcriptomes with histopathological morphology, presenting exciting new opportunities for biomolecular discovery. Using ST data, we construct a novel framework, GeneFlow, to map transcriptomics onto paired cellular images. By combining an attention-based RNA encoder with a conditional UNet guid

Haoran Yin, Yanqun Tang, Jun Xiong, Fan Liu

Integrated sensing and communication (ISAC) is a key feature of next-generation wireless networks, enabling a wide range of emerging applications such as vehicle-to-everything (V2X) and unmanned aerial vehicles (UAVs), which operate in high-mobility scenarios. Notably, the wireless channels within these applications typically exhibit severe delay and Doppler

Sutirtha Sengupta, Das Sujit, Arkaprabha Sarangi

We investigate the evolution of red supergiant (RSG) progenitors of core-collapse (CC) supernovae (SNe) with initial masses between $12-20~M_\odot$ focusing on the effects of enhanced mass loss due to pulsation-driven instabilities in their envelopes and subsequent dynamical ejections during advanced stages of nuclear burning. Using time-dependent mass loss

Koji Uryu, Shijun Yoshida, Eric Gourgoulhon, Charalampos Markakis

It is known that the mass of magnetized relativistic compact star is larger than that of non-magnetized one for the same equation of state and central density, albeit the excess of mass is sizable only if the magnetic fields are strong enough B~10^17-10^18G. Using our recently developed numerical code COCAL, we systematically compute such compact star soluti

Tanay Raghunandan Srinivasa, Suraj Kumar

This paper presents a physics-informed machine learning approach for synthesizing optimal feedback control policy for infinite-horizon optimal control problems by solving the Hamilton-Jacobi-Bellman (HJB) partial differential equation(PDE). The optimal control policy is derived analytically for affine dynamical systems with separable and strictly convex cont

Zixuan Hu, Yongxian Wei, Li Shen, Zhenyi Wang

Model inversion, which aims to reconstruct the original training data from pre-trained discriminative models, is especially useful when the original training data is unavailable due to privacy, usage rights, or size constraints. However, existing dense inversion methods attempt to reconstruct the entire image area, making them extremely inefficient when inve

Hoang Hiep Ly, Cong-Nhat Nguyen, Doan-Quang Tran, Quoc-Khanh Dang

Grasping objects with diverse mechanical properties, such as heavy, slippery, or fragile items, remains a significant challenge in robotics. Conventional grippers often rely on applying high normal forces, which can cause damage to objects. To address this limitation, we present a hybrid gripper finger that combines a rigid structural shell with a soft, infl

Mason Shipton, York Hay Ng, Aditya Khan, Phuong Hanh Hoang

The URIEL+ linguistic knowledge base supports multilingual research by encoding languages through geographic, genetic, and typological vectors. However, data sparsity (e.g. missing feature types, incomplete language entries, and limited genealogical coverage) remains prevalent. This limits the usefulness of URIEL+ in cross-lingual transfer, particularly for

Stanislav Selitskiy

Contemporary e-mail services have high availability expectations from the customers and are resource-strained because of the high-volume throughput and spam attacks. Deep Machine Learning architectures, which are resource hungry and require off-line processing due to the long processing times, are not acceptable at the front line filters. On the other hand,

Zongshun Zhang, Ibrahim Matta

Dynamic offloading of Machine Learning (ML) model partitions across different resource orchestration services, such as Function-as-a-Service (FaaS) and Infrastructure-as-a-Service (IaaS), can balance processing and transmission delays while minimizing costs of adaptive inference applications. However, prior work often overlooks real-world factors, such as Vi

Guozheng Zheng, Jian Guan, Mingjie Xie, Xuanjia Zhao

Cross-view geo-localization (CVGL) between drone and satellite imagery remains challenging due to severe viewpoint gaps and the presence of hard negatives, which are visually similar but geographically mismatched samples. Existing mining or reweighting strategies often use static weighting, which is sensitive to distribution shifts and prone to overemphasizi

Hao Jiao

A radiation field can be excited via parametric resonance when an oscillating axion field couples to the electromagnetic sector through a Chern-Simons interaction. As demonstrated in previous works, this mechanism can generate primordial magnetic fields shortly after recombination and provide sufficient ultraviolet radiation for the formation of direct colla

Guanchong Huang, Song Fang

Video identification attacks pose a significant privacy threat that can reveal videos that victims watch, which may disclose their hobbies, religious beliefs, political leanings, sexual orientation, and health status. Also, video watching history can be used for user profiling or advertising and may result in cyberbullying, discrimination, or blackmail. Exis

Xuan-Thuan Nguyen, Khac Nam Nguyen, Ngoc Duy Tran, Thi Thoa Mac

Multi-robot systems, particularly mobile manipulators, face challenges in control coordination and dynamic stability when working together. To address this issue, this study proposes MobiDock, a modular self-reconfigurable mobile manipulator system that allows two independent robots to physically connect and form a unified mobile bimanual platform. This proc

Shengze Xue, Yu Ren, Jiacheng Hong, Run Ni

Fast Fourier Transform-based (FFT) spectral oceans are widely adopted for their efficiency and large-scale realism, but they assume global stationarity and spatial homogeneity, making it difficult to represent non-uniform seas and near-field interactions (e.g., ships and floaters). In contrast, wave particles capture local wakes and ripples, yet are costly t

Junfan Li, Shizhong Liao, Zenglin Xu, Liqiang Nie

In this paper, we study the problem of online sparse linear regression (OSLR) where the algorithms are restricted to accessing only $k$ out of $d$ attributes per instance for prediction, which was proved to be NP-hard. Previous work gave polynomial-time algorithms assuming the data matrix satisfies the linear independence of features, the compatibility condi

Pouya Hamadanian, Pantea Karimi, Arash Nasr-Esfahany, Kimia Noorbakhsh

Can AI autonomously design mechanisms for computer systems on par with the creativity and reasoning of human experts? We present Glia, an AI architecture for networked systems design that uses large language models (LLMs) in a human-inspired multi-agent workflow. Each agent specializes in reasoning, experimentation, and analysis, collaborating through an eva

Gaoyuan Zhang, Gaolei Song, Boyuan Li, Zijian Li

From the perspective of hard decision fusion, we investigate Byzantine attacks in Reconfigurable Intelligent Surface (RIS)-enhanced and decode-and-forward relay-assisted Cooperative Spectrum Sensing (CSS) for mobile Cognitive Radio Networks (CRNs) in this paper. Specially, a RIS-enhanced and decode-and-forward relay-assisted CSS configuration is first constr

Jiaxin Yuan, Haizhao Yang, Maria Cameron

Fast and accurate simulation of dynamical systems is a fundamental challenge across scientific and engineering domains. Traditional numerical integrators often face a trade-off between accuracy and computational efficiency, while existing neural network-based approaches typically require training a separate model for each case. To overcome these limitations,

Zixuan Hu, Li Shen, Zhenyi Wang, Yongxian Wei

Harmful fine-tuning poses critical safety risks to fine-tuning-as-a-service for large language models. Existing defense strategies preemptively build robustness via attack simulation but suffer from fundamental limitations: (i) the infeasibility of extending attack simulations beyond bounded threat models due to the inherent difficulty of anticipating unknow

Mingyu Sung, Il-Min Kim, Sangseok Yun, Jae-Mo Kang

Diffusion models have emerged as state-of-the-art in image generation, but their practical deployment is hindered by the significant computational cost of their iterative denoising process. While existing caching techniques can accelerate inference, they often create a challenging trade-off between speed and fidelity, suffering from quality degradation and h

Tianlei Chen, Yuxiao Chen, Yang Li, Feifei Wang

Automated international classification of diseases (ICD) coding aims to assign multiple disease codes to clinical documents and plays a critical role in healthcare informatics. However, its performance is hindered by the extreme long-tail distribution of the ICD ontology, where a few common codes dominate while thousands of rare codes have very few examples.

Partha Ghose

Nelson's stochastic mechanics links quantum mechanics to an underlying Brownian motion with the identification $\hbar = m\sigma$. Ghose's interpolating equation introduces a continuous parameter $\lambda$ that suppresses the quantum potential $Q[\psi]$ and yields a smooth transition between quantum ($\lambda=0$) and classical ($\lambda=1$) regimes. In this s

Yucheng Xing, Jinxing Yin, Xiaodong Liu

Recently, diffusion models have shown their impressive ability in visual generation tasks. Besides static images, more and more research attentions have been drawn to the generation of realistic videos. The video generation not only has a higher requirement for the quality, but also brings a challenge in ensuring the video continuity. Among all the video gen

Jing Chang, Chang Liu, Jinbin Huang, Shuyuan Zheng

Automated data preparation pipeline construction is critical for machine learning success, yet existing methods suffer from two fundamental limitations: they treat pipeline construction as black-box optimization without quantifying individual operator contributions, and they struggle with the combinatorial explosion of the search space ($N^M$ configurations

Xiaozhi Li, Huijun Di, Jian Li, Feng Liu

Recent advances in 4D imaging radar have enabled robust perception in adverse weather, while camera sensors provide dense semantic information. Fusing the these complementary modalities has great potential for cost-effective 3D perception. However, most existing camera-radar fusion methods are limited to single-frame inputs, capturing only a partial view of

Yan Zhu, Qingyang Liu, Chang Guo, Tianlong Fan

Rumor propagation in social networks undermines social stability and public trust, calling for interventions that are both effective and resource-efficient. We develop a node-level, time-varying optimal intervention framework that allocates limited resources according to the evolving diffusion state. Unlike static, centrality-based heuristics, our approach d

Hankyeol Lee, Gawon Seo, Kyounggyu Lee, Dogun Kim

Vision-language models (VLMs) often struggle to generate accurate and detailed captions for high-resolution images since they are typically pre-trained on low-resolution inputs (e.g., 224x224 or 336x336 pixels). Downscaling high-resolution images to these dimensions may result in the loss of visual details and the omission of important objects. To address th

Jieshan Chen, Suyu Ma, Qinghua Lu, Sung Une Lee

Before deploying an AI system to replace an existing process, it must be compared with the incumbent to ensure improvement without added risk. Traditional evaluation relies on ground truth for both systems, but this is often unavailable due to delayed or unknowable outcomes, high costs, or incomplete data, especially for long-standing systems deemed safe by

Ruifei Huang, Hao Ning, Ze Zhong, Ye Qiu

Flare ribbons with parallel and circular morphologies are typically associated with different magnetic reconnection models, and the simultaneous observation of both types in a single event remains rare. Using multi-wavelength observations from a tandem of instruments, we present an M8.2-class flare that occurred on 2023 September 20, which produced quasi-par

Xiying Du, Yanjia Li, Xingxing Yu

Let $G$ be a graph and $x_1, x_2, \ldots, x_k$ be distinct vertices of $G$. We say $(G,x_1x_2\ldots x_k)$ has a $C_k$-minor or $G$ has a $C_k$-minor rooted at $x_1x_2\ldots x_k$, if there exist pairwise disjoint sets $X_1, X_2, \ldots, X_k\subseteq V(G)$, such that for all $i\in [k]$, $G[X_i]$ is connected, $x_i\in X_i$, and $G$ has an edge between $X_i$ and

Mitsuhiro Nishijima, Pierre-Louis Poirion, Akiko Takeda

In this paper, we propose a subgradient algorithm with a non-asymptotic convergence guarantee to solve copositive programming problems. The subproblem to be solved at each iteration is a standard quadratic programming problem, which is NP-hard in general. However, the proposed algorithm allows this subproblem to be solved inexactly. For a prescribed accuracy

Chuangqiang Hu, Xiuwu Zhu

We introduce two new types of towers of Drinfeld modular curves. These towers originate from a specific domain $\mathcal{A} $ and are analogous to the towers of rank-two Drinfeld modular curves over the polynomial ring. Specifically, the domain $\mathcal{A} $ corresponds to the projective line over the finite field $ \mathbb{F}_q $, equipped with an infinite

Yanfan Zhu, Juming Xiong, Ruining Deng, Yu Wang

The Banff Classification provides the global standard for evaluating renal transplant biopsies, yet its semi-quantitative nature, complex criteria, and inter-observer variability present significant challenges for computational replication. In this study, we explore the feasibility of approximating Banff lesion scores using existing deep learning models thro

Alexander Bakumenko, Aaron J. Masino, Janine Hoelscher

The translation of artificial intelligence (AI) systems into clinical practice requires bridging fundamental gaps between explainable AI theory, clinician expectations, and governance requirements. While conceptual frameworks define what constitutes explainable AI (XAI) and qualitative studies identify clinician needs, little practical guidance exists for de