December 2025 arXiv papers — page 110

Roji Pius

Feynman diagrams are the foremost tool in the perturbative study of quantum field theory. In gauge theories, the full potential of this tool is revealed when it is combined with the Slavanov-Taylor identities associated with the local gauge symmetry. Hence, it is desirable to have perturbative expansion of scattering amplitudes that combine the graphical nat

Chen Huang, Ronghui Hou

Connected and automated vehicles (CAVs) rely on wireless communication to exchange state information for distributed control, making communication delays a critical factor that can affect vehicle motion and degrade control performance, particularly in high-speed scenarios. To address these challenges in the complex environment of roundabout intersections, th

Jaeyoon Kim, Yoonki Cho, Sung-Eui Yoon

Visual Place Recognition (VPR) has advanced significantly with high-capacity foundation models like DINOv2, achieving remarkable performance. Nonetheless, their substantial computational cost makes deployment on resource-constrained devices impractical. In this paper, we introduce an efficient asymmetric VPR framework that incorporates a high-capacity galler

Zhentao Liang, Nees Jan van Eck, Xuehua Wu, Jin Mao

Effective science mapping relies on high-quality representations of scientific documents. As an important task in scientometrics and information studies, science mapping is often challenged by the complex and heterogeneous nature of citations. While previous studies have attempted to improve document representations by integrating citation and semantic infor

Dong Hui Han, Kyoung-Woong Moon, Kab-Jin Kim, Se Kwon Kim

The dynamics of a cycloidal spin structure driven by an AC magnetic field is theoretically studied in the weak-field limit. A specific model Hamiltonian describing the cycloidal spin structure in a ferromagnetic thin film is constructed, and its dynamics is analyzed using the collective-coordinate approach within the Lagrangian formalism. We demonstrate that

Shlomo Barak, George Salman

We present an adaptive geometry in which the yardstick co-deforms with space itself, formulated on cellular spaces where length is a count: distances are shortest cell-crossing counts. No cell shape, angles, or embedding are assumed; the framework is deliberately micro-agnostic. Curvature and deformation are inferred operationally by comparing a measured rad

Zhixuan Wen, Sheng Mao, Huiling Duan, Fan Feng

Curve-fold origami, composed of developable panels joined along a curved crease, exhibits rich dynamic behaviors relevant to metamaterials and soft robotic systems. Despite multiple approximated models, a comprehensive and exact dynamical theory for curve-fold origami remains absent, limiting the precise predictions of its dynamics, especially for those with

Bruno Cessac, Erwan Demairy, Jérôme Emonet, Evgenia Kartsaki

We developed Macular, a simulation platform with a graphical interface, designed to produce in silico experiment scenarios for the retina and the primary visual system. A scenario consists of generating a three-dimensional structure with interconnected layers, each layer corresponding to a type of 'cell' in the retina or visual cortex. The cells can correspo

Ruifan Chu, Anbang Wang, Xiuxiu Bai, Shuai Liu

In high-performance computing, hotspot GPU kernels are primary bottlenecks, and expert manual tuning is costly and hard to port. Large language model methods often assume kernels can be compiled and executed cheaply, which fails in large applications where full builds and runs are expensive. We present an end-to-end LLM framework with performance feedback th

Yuxiu Lu

We generalize the classical Fisher information metric on statistical models to $L^p$-metrics on various spaces of differential forms or group of diffeomorphisms. Using this new interpretation from information geometry, we derive several new results in geometry on group of diffeomorphisms, symplectic geometry and Teichm\"{u}ller theory. This includes geometry

Emilie Hein, Polin Yadak, Denise Hum, Jessica Hurless

Providing meaningful research experiences for undergraduate students is a well-recognized challenge, particularly at community colleges and teaching-focused institutions where resources are limited and faculty time is dedicated to instruction. To address this, the Summer Introduction to Research and Experimentation in Nuclear physics (SIREN) was developed as

Le Bin Ho

Quantum Chromodynamics (QCD) admits a topological $\bar{\theta}$ term that violates charge-parity ($CP$) symmetry, yet experiments indicate that $\bar{\theta}$ is extremely small. To investigate this problem in a controlled setting, we derive a Hamiltonian formulation of QCD through a $(1+1)$-dimensional Schwinger-model analogue. Fermionic and gauge degrees

Tapan K. Sasmal, Soumen Bera, Sabyasachi Pal, Soumen Mondal

The head--tail (HT) morphology of radio galaxies is seen for a class of radio sources where the primary lobes are being bent in the intercluster weather due to strong interactions between the radio jets and their respective intracluster medium. A systematic search has been carried out for new HT radio galaxies from the Very Large Array Faint Images of the Ra

Qingyuan Liu, Mo Zou, Hengbin Zhang, Dong Du

File systems are critical OS components that require constant evolution to support new hardware and emerging application needs. However, the traditional paradigm of developing features, fixing bugs, and maintaining the system incurs significant overhead, especially as systems grow in complexity. This paper proposes a new paradigm, generative file systems, wh

You-Wei Ding, Yen Chin Ong, Hao Xu

In a seminal paper, Abbott et al. analyzed the relationship between a particle's trajectory and the resolution of position measurements performed by an observer at fixed time intervals. They predicted that quantum paths exhibit a universal Hausdorff dimension that transitions from $d=2$ to $d=1$ as the momentum of the particle increases. However, although me

Stefan Vey, Christian Oliver Paschereit, David Greenblatt

Dimensionless frequency scaling laws for active separation control on flat-plate wings, using dielectric barrier discharge plasma actuators, were examined on the basis of maximum increases to lift coefficient, and compared with hovering insect wing-flapping frequencies. Data for a range of angles of attack ($24^\circ$ to $32^\circ$), Reynolds numbers (3,000

Thanh Nguyen Van Long, Lan Nguyen Tran, Le Bin Ho

Preparing ground states of strongly correlated quantum systems is a central goal in quantum simulation and optimization. The feedback-based quantum algorithm (FALQON) provides an attractive alternative to variational methods with a fully quantum feedback rule, but it fails in the presence of spectral degeneracies, where the feedback signal collapses and the

Tong Wei, Yijun Yang, Changhao Zhang, Junliang Xing

Multi-turn reinforcement learning (RL) for multi-modal agents built upon vision-language models (VLMs) is hampered by sparse rewards and long-horizon credit assignment. Recent methods densify the reward by querying a teacher that provides step-level feedback, e.g., Guided Thought Reinforcement (GTR) and On-Policy Distillation, but rely on costly, often privi

Tomohiro Okuma, Kei-ichi Watanabe, Ken-ichi Yoshida

Let $A$ be an excellent two-dimensional normal local ring containing an algebraically closed field and let $X\to \mathrm{Spec} (A)$ be a resolution of singularity. We prove a theorem giving a condition under which the dimension of the cohomology group of invertible sheaves on $X$ coincides with a natural lower bound. Applying this theorem, we establish upper

Daniel Veldman, Yue Wang

In this paper, a stochastic algorithm for the efficient simulation and optimal control of networked wave equations based on the random batch method is proposed and analyzed. The random approximation is constructed by dividing the time interval into subintervals and restricting the dynamics to a randomly chosen subnetwork during each of these subintervals. It

Xuwei Tan, Yao Ma, Xueru Zhang

Detecting fraud in financial transactions typically relies on tabular models that demand heavy feature engineering to handle high-dimensional data and offer limited interpretability, making it difficult for humans to understand predictions. Large Language Models (LLMs), in contrast, can produce human-readable explanations and facilitate feature analysis, pot

Hao Chen, Yiwei Wang, Songze Li

Concept erasure, which fine-tunes diffusion models to remove undesired or harmful visual concepts, has become a mainstream approach to mitigating unsafe or illegal image generation in text-to-image models.However, existing removal methods typically adopt a unidirectional erasure strategy by either suppressing the target concept or reinforcing safe alternativ

Milind Sarkar, Maria Giovanna Dainotti, Nikita S. Khatiya, Dhruv S. Bal

Gamma-ray bursts (GRBs) are among the most luminous explosions in the Universe and serve as powerful probes of the early cosmos. However, the rapid fading of their afterglows and the scarcity of spectroscopic measurements make photometric classification crucial for timely high-redshift identification. We present an ensemble machine learning framework for red

Taoran Sheng, Sathappan Muthiah, Atiq Islam, Jinming Feng

In e-commerce shopping, aligning search results with a buyer's immediate needs and preferences presents a significant challenge, particularly in adapting search results throughout the buyer's shopping journey as they move from the initial stages of browsing to making a purchase decision or shift from one intent to another. This study presents a systematic ap

C. Abinash Bhuyan, Anil K. Chaudhary, Kishore K. Madapu, P. Naveen Kumar

For designing an efficient terahertz (THz) emitter, the defect density of the semiconductors is smartly increased to reduce carrier lifetime, which subsequently lowers the overall power output of the semiconductor. To overcome this fundamental trade-off, this study presents a novel approach, by integrating a direct band gap 2D semiconductor such as monolayer

Fran Ilcic, Indrakshi Raychowdhury

Within the aim of understanding quantum chromodynamics through simulation, an increasingly studied approach is that of quantum computation and simulation. Challenges exist in encoding the minimal and physical degrees of freedom for a non-Abelian gauge theory and maintaining physical or gauge-invariant dynamics in a simulation. In this work, the Loop-String-H

Xiaoyu He, Yu Cai, Jin Jia, Canxi Huang

This work proposes Alada, an adaptive momentum method for stochastic optimization over large-scale matrices. Alada employs a rank-one factorization approach to estimate the second moment of gradients, where factors are updated alternatively to minimize the estimation error. Alada achieves sublinear memory overheads and can be readily extended to optimizing t

Jongwook Kim, Sangheon Yun, Sukjin Yoon

The canonical $O(N^2)$ Transformer remains the empirical performance frontier in sequence modeling, and its training can be further optimized by addressing geometric inefficiency. We propose an optimization framework that leverages an asymmetric projection to decompose the backward-pass gradients into parallel spans and orthogonal violations, while keeping t

Emir Aslandogan, Haci Ilhan, Burak Ahmet Ozden, Erdogan Aydin

Orthogonal time-frequency space (OTFS) modulation has emerged as a powerful wireless communication technology that is specifically designed to address the challenges of high-mobility scenarios and significant Doppler effects. Unlike conventional modulation schemes that operate in the time-frequency (TF) domain, OTFS projects signals to the delay-Doppler (DD)

Jie Qin, Jiancheng Huang, Limeng Qiao, Lin Ma

Multimodal large language models (MLLMs) play a pivotal role in advancing the quest for general artificial intelligence. However, achieving unified target for multimodal understanding and generation remains challenging due to optimization conflicts and performance trade-offs. To effectively enhance generative performance while preserving existing comprehensi

Tomoya Tanaka, Tomonori Ikeda, Ryo Yonemoto

This study presents the first comprehensive comparison of rule-based methods, traditional machine learning models, and deep learning models in radio wave sensing with frequency modulated continuous wave multiple input multiple output radar. We systematically evaluated five approaches in two indoor environments with distinct layouts: a rule-based connected co

Hongzhe Bi, Hengkai Tan, Shenghao Xie, Zeyuan Wang

While a general embodied agent must function as a unified system, current methods are built on isolated models for understanding, world modeling, and control. This fragmentation prevents unifying multimodal generative capabilities and hinders learning from large-scale, heterogeneous data. In this paper, we propose Motus, a unified latent action world model t

Sayan Kumar Pal

We present here an interesting non-relativistic limit, referred to as the Newton-Hooke (NH) limit, of the purely magnetic BTZ solution by starting from the Einstein-Maxwell system in the 2+1 dimensions. The Newton-Hooke limit is different from the Galilean limit in the sense that the former contains an additional parameter {\Lambda}, the cosmological constan

Roman Z. Morawski

The ubiquitous use of the tools of artificial intelligence (AI) in techno-science, in higher education and in other existing and potential fields of measurement application generates new challenges for teaching measurement science and technology (MST). The aim of this article is to encourage its readers to modernize their approach to teaching MST in a way as

Makoto Nagata, Yoshinori Takei

This paper proves that two differently defined rooted binary trees are isomorphic. The first tree is one associated to a version of Farey sequences where the vertices correspond to the open intervals formed by two successive terms in the sequence. The other tree has the vertices consisting of pairs of positive integers whose adjacency is defined by a simple

I. A. Seit-Dzhelil, A. Yu. Shvets

Non-ideal deterministic system "tank with liquid-electric motor" is studied. Two delay-approximation models are considered. Impact of the delay on the emergence, evolution and disappearance of regular and chaotic limit sets (attractors) of the system is investigated. The main dynamic characteristics of the system's steady-state regimes are computed and analy

Jun-Jin Peng

As a contribution towards the understanding for the field equations of diffeomorphism invariant theories of pure gravity, we demonstrate in great detail that the expression for the field equations of such theories can be derived within the perspective of the surface term coming from the variation of the Lagrangian. Specifically, starting with the surface ter

Teng Yin, Hao Zhang

In non-Hermitian physics, high-order exceptional points(HOEPs) with eigenvalues and eigenvectors coalesce are known for their enhanced sensitivity to perturbations. Typically, they exhibit eigenvalue splitting that scales as {\epsilon}^(1/n), which is referred to as the generic response. However, under certain conditions, a nongeneric response of HOEPs occur

Li Xiao

Since the formal introduction of its "dual-carbon" strategy in 2020, China has witnessed the concepts of green development and sustainability evolve from policy directives into a broad societal consensus. Within this transformative context, the Environmental, Social, and Governance (ESG) framework has emerged as a critical enabler, mutually reinforcing and s

Soohyung Lee, Ho Jun Jeong, JongMo Hwang, GwangUk Park

Beam position monitors (BPMs) are indispensable components of modern particle accelerators, providing real-time diagnostics to ensure precise beam control, stability, and quality. As accelerators such as the International Linear Collider (ILC) aim for nanometer-scale beam sizes at the interaction point, stringent requirements on position resolution arise. Sp

Mo Yang, Jing Yu, Necmiye Ozay

In many multi-agent systems, communication is limited by bandwidth, latency, and energy constraints. Designing controllers that achieve coordination and safety with minimal communication is critical for scalable and reliable deployment. This paper presents a method for designing controllers that minimize inter-agent communication in multi-agent systems while

Jia-Jun Ma, Congling Qiu, Zhiwei Yun, JiaLiang Zou

In this paper, we obtain an explicit formula for the theta correspondence of unipotent principal-series representations between an even orthogonal and a symplectic group or between general linear groups over a finite field. The formula is in terms of the Springer correspondence. Along the way we prove general results about module categories of Hecke categori

Cheeun Hong, German Barquero, Fadime Sener, Markos Georgopoulos

Instructional video generation is an emerging task that aims to synthesize coherent demonstrations of procedural activities from textual descriptions. Such capability has broad implications for content creation, education, and human-AI interaction, yet existing video diffusion models struggle to maintain temporal consistency and controllability across long s

Tomoya Tanaka, Tomonori Ikeda, Ryo Yonemoto

This study presents the first comprehensive evaluation of spatial generalization techniques, which are essential for the practical deployment of deep learning-based radio-frequency (RF) sensing. Focusing on people counting in indoor environments using frequency-modulated continuous-wave (FMCW) multiple-input multiple-output (MIMO) radar, we systematically in

K. U. Can, J. A. Crawford, R. Horsley, J. J. McKee

At large momentum transfer, it becomes increasingly difficult to access the form factor of the pion $F_\pi(Q^2)$ using lattice QCD simulations. Two of the limiting factors include the increased computational cost of adding more statistics to overcome gauge noise, as well as suppressed overlap with the ground state of the boosted pion. Here we apply two noise

Xinjie Li, Zhimin Chen, Rui Zhao, Florian Schiffers

Recent unified models for joint understanding and generation have significantly advanced visual generation capabilities. However, their focus on conventional tasks like text-to-video generation has left the temporal reasoning potential of unified models largely underexplored. To address this gap, we introduce Next Scene Prediction (NSP), a new task that push

Haoyu Wang, Lei Zhang, Wenrui Liu, Dengyang Jiang

Given the inherently costly and time-intensive nature of pixel-level annotation, the generation of synthetic datasets comprising sufficiently diverse synthetic images paired with ground-truth pixel-level annotations has garnered increasing attention recently for training high-performance semantic segmentation models. However, existing methods necessitate to

Katsuya Shimabukuro, Kosaku Horinaga, Kazumo Wakabayashi, Hikaru Emoto

The transduction of force into motion for microswimmers at intermediate Reynolds numbers ($Re \sim 1$), where inertia becomes relevant, is a fundamental problem in active matter. Using the multicellular alga \textit{Volvox} as a model physical system, we perform the first direct measurements that deconvolve a swimmer's inertial impact force from its motor's

Menglu Li, Majd Alber, Ramtin Asgarianamiri, Lian Zhao

Detecting partial deepfake speech is challenging because manipulations occur only in short regions while the surrounding audio remains authentic. However, existing detection methods are fundamentally limited by the quality of available datasets, many of which rely on outdated synthesis systems and generation procedures that introduce dataset-specific artifac

Zeren Simon Wang, Yu Zhang

In recent years, a number of experiments dedicated to searches for long-lived particles (LLPs) have been proposed, approved, or have entered operation. While the sensitivities of these experiments to various LLP scenarios have been extensively studied, key aspects--such as detector geometries, background estimates, and projected operational durations--for se

Hassan Iftikhar, Rizwan Ahmad, Arunark Kolipaka

The Multimodal Direct Inversion (MMDI) algorithm is widely used in Magnetic Resonance Elastography (MRE) to estimate tissue shear stiffness. However, MMDI relies on the Helmholtz equation, which assumes wave propagation in a uniform, homogeneous, and infinite medium. Furthermore, the use of the Laplacian operator makes MMDI highly sensitive to noise, which c

Xi Luo, Shixin Xu, Ying Xie, JianZhong Hu

Accurate medical image analysis can greatly assist clinical diagnosis, but its effectiveness relies on high-quality expert annotations Obtaining pixel-level labels for medical images, particularly fundus images, remains costly and time-consuming. Meanwhile, despite the success of deep learning in medical imaging, the lack of interpretability limits its clini

Md Nahid Hasan Shuvo, Moinul Hossain

Connected autonomous vehicles (CAVs) rely on vision-based deep neural networks (DNNs) and low-latency (Vehicle-to-Everything) V2X communication to navigate safely and efficiently. Despite their advances, these systems remain vulnerable to physical adversarial attacks. In this paper, we introduce PHANTOM (PHysical ANamorphic Threats Obstructing connected vehi

Nikolai Goncharov, James L. Gray, Donald G. Dansereau

Object tracking is an important step in robotics and reautonomous driving pipelines, which has to generalize to previously unseen and complex objects. Existing high-performing methods often rely on pre-captured object views to build explicit reference models, which restricts them to a fixed set of known objects. However, such reference models can struggle wi

Yifan Pu, Yizeng Han, Zhiwei Tang, Jiasheng Tang

Diffusion distillation has dramatically accelerated class-conditional image synthesis, but its applicability to open-ended text-to-image (T2I) generation is still unclear. We present the first systematic study that adapts and compares state-of-the-art distillation techniques on a strong T2I teacher model, FLUX.1-lite. By casting existing methods into a unifi

Apala Chakrabarti

Designing sustainable medical devices requires balancing environmental, economic, and social demands, yet trade-offs across these pillars are difficult to identify using manual assessment alone. Current methods depend heavily on expert judgment, lack standardisation, and struggle to integrate diverse lifecycle data, which leads to overlooked conflicts and in

Muhammad Sarwar, Muhammad Rizwan, Mubushra Aziz, Abdul Rehman Sudais

This comprehensive literature review examines the emerging applications of Large Language Models (LLMs) in power system engineering. Through a systematic analysis of recent research published between 2020 and 2025, we explore how LLMs are being integrated into various aspects of power system operations, planning, and management. The review covers key applica

Min Lu, Hemant Ishwaran

Out-of-distribution (OOD) detection is essential for determining when a supervised model encounters inputs that differ meaningfully from its training distribution. While widely studied in classification, OOD detection for regression and survival analysis remains limited due to the absence of discrete labels and the challenge of quantifying predictive uncerta

Shoot Koebisu

We study zero-divisors in the $16$-dimensional sedenion algebra from the viewpoint of the determinant of left multiplication. We show that this determinant admits a canonical factorization into the square of a quartic polynomial, obtained via a $G_2$-invariant reduction to a quaternionic normal form and an explicit block computation. The quartic factor recov

Taero Kim, Hoyoon Byun, Youngjun Choi, Sungrae Park

Expanding pre-trained language models offers a practical way to increase capacity without training larger models from scratch. Depth Up-Scaling (DUS) does so by duplicating Transformer blocks and inserting them into a pre-trained backbone. This process also duplicates FFN-heavy blocks, increasing parameter and compute cost while adding capacity through a blo

Genki Kusano, Kenya Abe, Kunihiro Takeoka

Recommender systems usually rely on large-scale interaction data to learn from users' past behaviors and make accurate predictions. However, real-world applications often face situations where no training data is available, such as when launching new services or handling entirely new users. In such cases, conventional approaches cannot be applied. This study

Runhua Zhang, Ruyuan Wan, Jiaqi Li, Daye Kang

Feminist HCI has been rapidly developing in East Asian contexts in recent years. The region's unique cultural and political backgrounds have contributed valuable, situated knowledge, revealing topics such as localized digital feminism practices, or women's complex navigation among social expectations. However, the very factors that ground these perspectives

Leonardo Bohac

Quantum random access memory (QRAM) is a central primitive for coherent data access in quantum algorithms, yet it remains controversial in practice because the wall-clock cost of "one lookup" can hide routing depth, control overhead, and geometric constraints. We present a universal QRAM construction (U-QRAM) in which the database is a physical memory regist

A. Sanna, G. Illiano, M. C. Baglio, D. M. Russell

MAXI J1957+032 is an accreting millisecond X-ray pulsar that shows brief, recurrent outbursts in an ultra-compact ~1 h orbit. We characterise the 2025 outburst using X-ray timing and spectroscopy from XMM-Newton and Swift (and a late-time NuSTAR observation), together with contemporaneous optical photometry from LCO, and compare the spin frequency with the 2

Payel Shee, Tanaya Halder, Chia-Jung Yang, Nainish Tickoo

The interplay between magnetic order, charge dynamics, and crystal field excitations underpins the emergent ground states of rare-earth intermetallics. Using time-domain terahertz spectroscopy, we probe this coupling in PrSi, a metallic ferromagnet. The optical response exhibits pronounced Drude-Smith behavior over a broad temperature range, indicating persi

Ying-ze Shan, Wei-Hua Lei, Hao-Tian Lan, Shao-yu Fu

Fast radio bursts (FRBs) are bright, millisecond-duration radio bursts with poorly known origins. Most FRB sources are detected only once, while some are repeaters. Variation patterns observed in the rotation measure (RM) of some repeaters -- indicate that the local magneto-ionic environments of these FRB sources are highly dynamic. It has been suggested tha

Ezieddin Elmahjub

Generative AI (GenAI) outputs are not copyrightable. This article argues why. We bypass conventional doctrinal analysis that focuses on black letter law notions of originality and authorship to re-evaluate copyright's foundational philosophy. GenAI fundamentally severs the direct human creative link to expressive form. Traditional theories utilitarian incent

Boyuan Ning, Yasutaka Shimizu

We study a twice-differentiable transformation applied to a CKLS-type short-rate model with linear drift and power-type diffusion. The transformation yields a new process whose diffusion component has a square-root structure and whose drift becomes nonlinear. A critical reassessment of earlier studies using similar transformations reveals fundamental errors

Guillermo A. Castillo, Himanshu Lodha, Ayonga Hereid

This work introduces a hierarchical strategy for terrain-aware bipedal locomotion that integrates reduced-dimensional perceptual representations to enhance reinforcement learning (RL)-based high-level (HL) policies for real-time gait generation. Unlike end-to-end approaches, our framework leverages latent terrain encodings via a Convolutional Variational Aut

Apala Chakrabarti

Medical devices improve healthcare outcomes but often involve sustainability conflicts across environmental, economic, and social pillars. Existing approaches typically prioritize one or two pillars and lack a unified framework to assess cross-domain conflicts. This paper presents a structured framework to identify and quantify sustainability conflicts in me

Soobin Cho, Panki Kim, Renming Song, Zoran Vondraček

In this paper, we study two types of purely discontinuous symmetric Markov processes $X$ in bounded smooth subsets of $\mathbb R^d$: conservative processes and processes killed either upon approaching the boundary of the set or by a killing potential $\kappa$. The jump kernel of $X$ is of the form $J(x,y)={\cal B}(x,y)|x-y|^{-d-\alpha}$, $\alpha\in (0,2)$, w

Yuseon Choi, Sangjin Kim, Jungjun Oh, Gwangtae Park

MoE models offer efficient scaling through conditional computation, but their large parameter size and expensive expert offloading make on-device deployment challenging. Existing acceleration techniques such as prefetching or expert clustering often increase energy usage or reduce expert diversity. We present SliceMoE, an energy-efficient MoE inference frame

Hailee Carter

As brain computer interfaces (BCIs) transition from experimental medical systems to consumer and military adjacent technologies, they introduce a novel security domain in which the human nervous system becomes a networked and contestable substrate. Existing frameworks for cybersecurity, biomedical safety, and data protection were not designed to address adve

Yilei Zhang, Yun Wei, Aritra Guha, XuanLong Nguyen

Mixture models are widely used in modeling heterogeneous data populations. A standard approach of mixture modeling assumes that the mixture component takes a parametric kernel form. In many applications, making parametric assumptions on the latent subpopulation distributions may be unrealistic, which motivates the need for nonparametric modeling of the mixtu

Amin Jalal Aghdasian, Farzaneh Abdollahi, Ali Kamali Iglie

This paper proposes two new algorithms for the lane keeping system (LKS) in autonomous vehicles (AVs) operating under snowy road conditions. These algorithms use deep reinforcement learning (DRL) to handle uncertainties and slippage. They include Action-Robust Recurrent Deep Deterministic Policy Gradient (AR-RDPG) and end-to-end Action-Robust convolutional n

Takayuki Hibi, Seyed Amin Seyed Fakhari

A lattice polytope $\mathcal{P} \subset \mathbb{R}^n$ of dimension $n$ is called level* if (i) $\mathcal{P}$ is normal, (ii) $(\mathcal{P} \setminus \partial \mathcal{P}) \cap \mathbb{Z}^n \neq \emptyset$ and (iii) for each $N = 2,3, \ldots$ and for each $\textbf{a} \in N(\mathcal{P} \setminus \partial \mathcal{P}) \cap \mathbb{Z}^n$, there is $\textbf{a}_0

Xiaoshan Huang, Maria Renee Meza, Sol Bin Yun, Brenna Mockler

The early time emission in tidal disruption events (TDEs) originates from both accretion and shocks, producing photons that eventually emerge from an inhomogeneous photosphere. We model disk formation following debris stream self-intersection in a TDE using three-dimensional, frequency-integrated and multi-group radiation hydrodynamic simulations. We find a

Jiayin Lu, Ying Jiang, Yumeng He, Yin Yang

Voronoi diagrams naturally produce convex, watertight, and topologically consistent cells, making them an appealing representation for 3D shape reconstruction. However, standard differentiable Voronoi approaches typically optimize generator positions in stable configurations, which can lead to locally uneven surface geometry. We present VoroLight, a differen

Moka Nishigaki, Kimihiko Nakajima, Masami Ouchi, Peter Behroozi

We present the statistics of line ratios and direct Te-based metallicities from JWST medium-resolution spectra of 292 galaxies at z=2-10, combining DREAMS observations with those of JADES and CEERS. To remove systematics caused by stellar mass (M*) and star formation rate (SFR), we construct stacked spectra binned by redshift within fixed M* and SFR ranges,

Fei Guo, Fangxia Wang

The diminished Sombor index $(DSO)$ of a graph $G$, introduced by Rajathagiri, is defined as $$DSO(G)=\sum_{uv\in E}\frac{\sqrt{d_u^2+d_v^2}}{d_u+d_v},$$ where $d_u$ and $d_v$ are the degrees of vertices $u$ and $v$. A graph $G$ is a molecular graph if $d_G(u)\leq 4$ for all $u\in V(G)$. In this paper, we examine the chemical applicability of the $DSO$ index

Ziheng Qin, Yuheng Ji, Renshuai Tao, Yuxuan Tian

The pursuit of a universal AI-generated image (AIGI) detector often relies on aggregating data from numerous generators to improve generalization. However, this paper identifies a paradoxical phenomenon we term the Benefit then Conflict dilemma, where detector performance stagnates and eventually degrades as source diversity expands. Our systematic analysis,

Jonathan Wenshøj, Tong Chen, Bob Pepin, Raghavendra Selvan

While joint pruning--quantization is theoretically superior to sequential application, current joint methods rely on auxiliary procedures outside the training loop for finding compression parameters. This reliance adds engineering complexity and hyperparameter tuning, while also lacking a direct data-driven gradient signal, which might result in sub-optimal

Joyjit Roy, Samaresh Kumar Singh

Financial sentiment analysis enhances market understanding. However, standard Natural Language Processing (NLP) approaches encounter significant challenges when applied to small datasets. This study presents a comparative evaluation of embedding-based techniques for financial news sentiment classification in resource-constrained environments. Word2Vec, GloVe

Tingyang Chen, Cong Fu, Jiahua Wu, Haotian Wu

Vector Similarity Search (VSS) in high-dimensional spaces is rapidly emerging as core functionality in next-generation database systems for numerous data-intensive services -- from embedding lookups in large language models (LLMs), to semantic information retrieval and recommendation engines. Current benchmarks, however, evaluate VSS primarily on the recall-

Christopher L. Rogers

We solve the differentiation problem for Lie $\infty$-groups. Our approach builds on a classical version of Cartier duality which canonically identifies the Hopf algebra of point distributions supported at the identity of a Lie group with the universal enveloping algebra of its Lie algebra. Hence, for Lie $\infty$-groups, we consider simplicial coalgebras of

Chee Heng Tan, Huiying Zheng, Jing Wang, Zhuoyi Lin

With the advent of large language models (LLMs), the landscape of recommender systems is undergoing a significant transformation. Traditionally, user reviews have served as a critical source of rich, contextual information for enhancing recommendation quality. However, as LLMs demonstrate an unprecedented ability to understand and generate human-like text, t

Shengling Qin, Hao Yu, Chenxin Wu, Zheng Li

This paper presents VLCache, a cache reuse framework that exploits both Key-Value (KV) cache and encoder cache from prior multimodal inputs to eliminate costly recomputation when the same multimodal inputs recur. Unlike previous heuristic approaches, we formally identify the cumulative reuse error effect and demonstrate how to minimize the non-prefix cache r

Marcus Ma, Cole Johnson, Nolan Bridges, Jackson Trager

Third-party annotation is the status quo for labeling text, but egocentric information such as sentiment and belief can at best only be approximated by a third-person proxy. We introduce author labeling, an annotation technique where the writer of the document itself annotates the data at the moment of creation. We collaborate with a commercial chatbot with

Chunyu Zou

Cryogenic electron microscopy (Cryo-EM) has become an essential tool for capturing high-resolution biological structures. Despite its advantage in visualizations, the large storage size of Cryo-EM data file poses significant challenges for researchers and educators. This paper investigates the application of deep learning, specifically implicit neural repres

Sixtus Dakurah

Identifying and comparing topological features, particularly cycles, across different topological objects remains a fundamental challenge in persistent homology and topological data analysis. This work introduces a novel framework for constructing cycle communities through two complementary approaches. First, a dendrogram-based methodology leverages merge-tr

Jun Jiang

In this paper, we study deformations of crossed homomorphisms on Lie groups by means of the cohomology which controls them. Using the Moser type argument, we obtain several rigidity results of crossed homomorphisms on Lie groups. We further investigate the relationship between the cohomology of crossed homomorphisms on Lie groups and that on Lie algebras. Fi

Zhongda Chu, Fei Teng

To address system frequency challenges driven by the integration of renewable generation, advanced control strategies are designed at the device level to provide effective frequency support following disturbances. However, typically relying on energy-based performance metrics, these methods cannot guarantee the system frequency constraints such as frequency

Andrei Zlotchevski, Linan Chen

The unbalanced Schr\"odinger bridge problem (uSBP) seeks to interpolate between a probability measure $\rho_0$ and a sub-probability measure $\rho_T$ while minimizing KL divergence to a reference measure $\mathbf{R}$ on a path space. In this work, we investigate the case where $\mathbf{R}$ is the path measure of a diffusion process with killing, which we int

Paola Di Maio

The intersection of artificial intelligence (AI) and digital forensics (DF) is becoming increasingly complex, ubiquitous, and pervasive, with overlapping techniques and technologies being adopted in all types of scientific and technical inquiry. Despite incredible advances, forensic sciences are not exempt from errors and remain vulnerable to fallibility. To

Andrew Francis, Charles Semple

Normal networks are an important class of phylogenetic networks that have compelling mathematical properties which align with intuition about inference from genetic data. While tools enabling widespread use of phylogenetic networks in the biological literature are still under mathematical, statistical, and computational development, many such results are bei

Yuetao Chen, Gaiqing Chen, Jin Wang, Qiang Ma

Recent advances in quantum optics have highlighted the critical role of spatial propagation in controlling the quantum coherence of light beams. However, the evolution of quantum coherence for light beams undergoing fundamental optical processes at dielectric interfaces remains unexplored. Furthermore, manipulating multiphoton correlations typically requires

Weizhou Shen, Ziyi Yang, Chenliang Li, Zhiyuan Lu

We introduce QwenLong-L1.5, a model that achieves superior long-context reasoning capabilities through systematic post-training innovations. The key technical breakthroughs of QwenLong-L1.5 are as follows: (1) Long-Context Data Synthesis Pipeline: We develop a systematic synthesis framework that generates challenging reasoning tasks requiring multi-hop groun

Thayssa Rocha, Luciano Teran, Marcelle Mota, Cleidson de Souza

The increasing adoption of remote and hybrid work modalities in the technology sector has brought new opportunities and challenges for the inclusion of people with disabilities (PWD) in software development teams (SDT). This study investigates how remote work affects PWDs' experience in mixed-ability SDT, focusing on the unique challenges and strategies that

Yupeng Li, Mingyue Cheng, Yucong Luo, Yitong Zhou

Multi-behavior sequential recommendation aims to capture users' dynamic interests by modeling diverse types of user interactions over time. Although several studies have explored this setting, the recommendation performance remains suboptimal, mainly due to two fundamental challenges: the heterogeneity of user behaviors and data sparsity. To address these ch

Luan Thanh Trinh, Kenji Doi, Atsuki Osanai

Diffusion models have emerged as the leading approach for style transfer, yet they struggle with photo-realistic transfers, often producing painting-like results or missing detailed stylistic elements. Current methods inadequately address unwanted influence from original content styles and style reference content features. We introduce SCAdapter, a novel tec