October 2025 arXiv papers — page 155

Tom Kimpson, Nicholas J. O'Neill, Patrick M. Meyers, Andrew Melatos

Argus is a high-performance Python package for detecting and characterising nanohertz gravitational waves in pulsar timing array data. The package provides a complete Bayesian inference framework based on state-space models, using Kalman filtering for efficient likelihood evaluation. Argus leverages JAX for just-in-time compilation, GPU acceleration, and aut

Lingyue Fu, Haowei Yuan, Datong Chen, Xinyi Dai

In programming education, Debugging and Teaching (DT) task is a common scenario where students receive assistance in correcting their erroneous code. The task involves multiple inputs, including erroneous code, error messages, reference solutions, and the question description, with the goal of generating modification suggestions to the erroneous code. Howeve

Mainak Dutta, Partha Nandi, Bibhas Ranjan Majhi

A central challenge in probing the quantum nature of gravity is to distinguish effects that are genuinely quantum from those that can be explained classically. In this work, we study how quantized gravitational waves interact with thermal quantum systems, modeled as harmonic oscillators. We show that, unlike classical waves, quantized gravitons generate enta

Jing Li, Bowei Guo, Xinqi Xie, Kuo-Ping Chang

The empirical results have shown that firstly, with one-week holding period and reinvesting, for SSE Composite Index stocks, the highest p-ratio investment strategy produces the largest annualized rate of return; and for NYSE Composite Index stocks, all the three strategies with both one-week and one-month periods generate negative returns. Secondly, with no

Yuan Tian, Min Zhou, Yitong Chen, Fang Li

Patient face images provide a convenient mean for evaluating eye diseases, while also raising privacy concerns. Here, we introduce ROFI, a deep learning-based privacy protection framework for ophthalmology. Using weakly supervised learning and neural identity translation, ROFI anonymizes facial features while retaining disease features (over 98\% accuracy, $

Huayi Wang, Wentao Zhang, Runyi Yu, Tao Huang

Deploying humanoid robots to interact with real-world environments--such as carrying objects or sitting on chairs--requires generalizable, lifelike motions and robust scene perception. Although prior approaches have advanced each capability individually, combining them in a unified system is still an ongoing challenge. In this work, we present a physical-wor

Han Cui, Jingbo Liu

Conditional sampling is a fundamental task in Bayesian statistics and generative modeling. Consider the problem of sampling from the posterior distribution $P_{X|Y=y^*}$ for some observation $y^*$, where the likelihood $P_{Y|X}$ is known, and we are given $n$ i.i.d. samples $D=\{X_i\}_{i=1}^n$ drawn from an unknown prior distribution $\pi_X$. Suppose that $f

Kang Zhou

We extend the recently discovered phenomenon of hidden zeros to tree amplitudes for Yang-Mills (YM) and general relativity (GR) theories with higher-derivative interactions. This includes gluon amplitudes with a single insertion of the local $F^3$ operator, as well as graviton amplitudes at sub-leading and sub-sub-leading orders in the low-energy expansion o

José Cantarero, Bernardo Villarreal

We express the set of representations from a cyclic $p$-group to a connected $p$-compact group in terms of the associated reflection group and compute its cardinality for each exotic $p$-compact group.

Xinyu Luo, Jie Liu, Kecheng Chen, Junyi Yang

Real-world deployment often exposes models to distribution shifts, making test-time adaptation (TTA) critical for robustness. Yet most TTA methods are unfriendly to edge deployment, as they rely on backpropagation, activation buffering, or test-time mini-batches, leading to high latency and memory overhead. We propose \textbf{ELaTTA} (\textit{Efficient Laten

Weimin Jiang, Juan Li, Yan Shen

This paper investigates $L^{1}$ solutions for mean-field backward stochastic differential equations (MFBSDEs) under different weak assumptions in both one-dimensional and multi-dimensional settings, whose generator $f(\omega,t,y,z,\mu)$ depends not only on the solution process $(Y,Z)$ but also on the law of $(Y,Z)$. In the one-dimensional case where $f$ depe

Alin Fan, Hanqing Li, Sihan Lu, Jingsong Yuan

Modern industrial recommendation systems improve recommendation performance by integrating multimodal representations from pre-trained models into ID-based Click-Through Rate (CTR) prediction frameworks. However, existing approaches typically adopt modality-centric modeling strategies that process ID-based and multimodal embeddings independently, failing to

Carlo Brunetta, Amit Chaudhary, Stefano Galatolo, Massimiliano Sala

Dynamically distributed inflation is a common mechanism used to guide a blockchain's staking rate towards a desired equilibrium between network security and token liquidity. However, the high sensitivity of the annual percentage yield to changes in the staking rate, coupled with the inherent feedback delays in staker responses, can induce undesirable oscilla

Isabella Graßl, Benedikt Fein, Gordon Fraser

Gender stereotypes in introductory programming courses often go unnoticed, yet they can negatively influence young learners' interest and learning, particularly under-represented groups such as girls. Popular tutorials on block-based programming with Scratch may unintentionally reinforce biases through character choices, narrative framing, or activity types.

Chang Liu, Henghui Ding, Kaining Ying, Lingyi Hong

This report presents an overview of the 7th Large-scale Video Object Segmentation (LSVOS) Challenge held in conjunction with ICCV 2025. Besides the two traditional tracks of LSVOS that jointly target robustness in realistic video scenarios: Classic VOS (VOS), and Referring VOS (RVOS), the 2025 edition features a newly introduced track, Complex VOS (MOSEv2).

Yujie Zhao, Lanxiang Hu, Yang Wang, Minmin Hou

Multi-agent systems (MAS) and reinforcement learning (RL) are widely used to enhance the agentic capabilities of large language models (LLMs). MAS improves task performance through role-based orchestration, while RL uses environmental rewards to learn stronger policies, such as GRPO-style optimization. However, applying on-policy RL to MAS remains underexplo

A. Dudko, S. Favorov

We prove that each discrete set in the Euclidean space that has bounded changes under every translation is a bounded perturbation of a square lattice, i.e., a uniformly spread set in the sense of Laszkovich. In particular, the support of every Fourier quasicrystal with unit masses is uniformly spread.

Jian Wang, Xiaofei Xie, Qiang Hu, Shangqing Liu

Automated Program Repair (APR) plays a critical role in enhancing the quality and reliability of software systems. While substantial progress has been made in Java-based APR, largely facilitated by benchmarks like Defects4J, there remains a significant gap in research on C/C++ program repair, despite the widespread use of C/C++ and the prevalence of associat

I Chiu, Yu-Tung Liu, Kuan-Chen Wang, Hung-Yu Wei

Photoplethysmography (PPG) is widely used in wearable health monitoring, but its reliability is often degraded by noise and motion artifacts, limiting downstream applications such as heart rate (HR) estimation. This paper presents a deep learning framework for PPG denoising with an emphasis on preserving physiological information. In this framework, we propo

Youngrok Park, Hojung Jung, Sangmin Bae, Se-Young Yun

Diffusion models have achieved remarkable success as generative models. However, even a well-trained model can accumulate errors throughout the generation process. These errors become particularly problematic when arbitrary guidance is applied to steer samples toward desired properties, which often breaks sample fidelity. In this paper, we propose a general

Runze Xia, Yupeng Ji, Yuxi Zhou, Haodong Liu

Query-service relevance prediction in e-commerce search systems faces strict latency requirements that prevent the direct application of Large Language Models (LLMs). To bridge this gap, we propose a two-stage reasoning distillation framework to transfer reasoning capabilities from a powerful teacher LLM to a lightweight, deployment-friendly student model. I

Na-Na Zhang, Chao-Yi Wu, Ming Li, Wei-Xuan Cao

The preservation of quantum coherence is besieged by a fundamental dogma: its revival necessitates non-Markovian memory effects from structured environments. This paradigm has constrained quantum control strategies and obscured simpler paths to coherence protection. Here, we shatter this belief by demonstrating unambiguous coherence revival even in strictly

JiSun Huh, Jang Soo Kim, Christian Krattenthaler, Soichi Okada

A Littlewood identity is an identity equating a sum of Schur functions with an infinite product. A bounded Littlewood identity is one where the sum is taken over the partitions with a bounded number of rows or columns. The price to pay is that the infinite product has to be replaced by a determinant. The focus of this article is on refinements of such bounde

Maurizio Palesi, Enrico Russo, Giuseppe Ascia, Hamaad Rafique

The scalability of quantum computing is constrained by the physical and architectural limitations of monolithic quantum processors. Modular multi-core quantum architectures, which interconnect multiple quantum cores (QCs) via classical and quantum-coherent links, offer a promising alternative to address these challenges. However, transitioning to a modular a

Qinglin Zhu, Yizhen Yao, Runcong Zhao, Yanzheng Xiang

Autoregressive (AR) models remain the standard for natural language generation but still suffer from high latency due to strictly sequential decoding. Recent diffusion-inspired approaches, such as LlaDA and Dream, mitigate this by generating in parallel, yet they suffer from two core limitations: information loss, as predictive distributions for non-finalize

Jong In Han, Jeong-Hoon Ju, Yeongrak Kim

We show that the border rank of the $4 \times 4$ determinant tensor is at least $12$ over $\mathbb{C}$, using the fixed ideal theorem introduced by Buczyńska-Buczyński and the method by Conner-Harper-Landsberg. Together with the known upper bound, this implies that the border rank is exactly $12$.

Yang Hou, Minggu Wang, Jianjun Zhao

Recent advancements in text-guided diffusion models have shown promise for general image editing via inversion techniques, but often struggle to maintain ID and structural consistency in real face editing tasks. To address this limitation, we propose a zero-shot face editing method based on ID-Attribute Decoupled Inversion. Specifically, we decompose the fac

Sonakshi Dua, Gonzalo Mateos, Sundeep Prabhakar Chepuri

Trustworthy decision making in networked, dynamic environments calls for innovative uncertainty quantification substrates in predictive models for graph time series. Existing conformal prediction (CP) methods have been applied separately to multivariate time series and static graphs, but they either ignore the underlying graph topology or neglect temporal dy

Onyedikachi J. Okeke, Uloma E. Nelson, Chukwudi Nwaogu, Olumide O. Oladoyin

This study applies Multiscale Geographically Weighted Regression (MGWR) to examine the spatial determinants of household wealth in Bernalillo County, New Mexico. The model incorporates sociodemographic, environmental, and proximity-based variables to evaluate how locational suitability influences economic outcomes. Key factors considered include income, home

Xiaxia Guan, Xian'an Jin, Weiling Yang

The Tutte polynomial is a significant invariant of graphs and matroids. It is well-known that it has three equivalent definitions: bases expansion, rank generating function, and deletion-contraction formula. The polymatroid Tutte polynomial $\mathscr{T}_{P}$ generalizes the Tutte polynomial from matroids to polymatroids $P$. In \emph{[Adv. Math. 402 (2022) 1

Yicheng Guang, Pietro Zanotta, Kai Zhou, Yueqi Chen

Classical program analysis techniques, such as abstract interpretation and symbolic execution, are essential for ensuring software correctness, optimizing performance, and enabling compiler optimizations. However, these techniques face computational limitations when analyzing programs with large or exponential state spaces, limiting their effectiveness in en

David Gobrecht

Traditionally, the condensation sequence of circumstellar dust is predicted based on the thermodynamic stabilities of specific condensates in the macroscopic bulk phase. However, at the (sub-)nanometer scale clusters with non-crystalline structures and significantly different properties are energetically favoured. For this reason, we study the thermodynamic

Mohd. Meraj Khan, Sumesh P. Thampi, Anubhab Roy

In this work, the lattice Boltzmann method (LBM) is assessed as a time-domain numerical approach for electromagnetic wave scattering. Owing to its explicit formulation and suitability for parallel computation on structured grids, LBM provides an alternative framework for solving Maxwell's equations. The formulation is first validated using canonical benchmar

Shiyao Zhang, Liwei Deng, Shuyu Zhang, Weijie Yuan

In future intelligent transportation systems, autonomous cooperative planning (ACP), becomes a promising technique to increase the effectiveness and security of multi-vehicle interactions. However, multiple uncertainties cannot be fully addressed for existing ACP strategies, e.g. perception, planning, and communication uncertainties. To address these, a nove

Wenya Xie, Qingying Xiao, Yu Zheng, Xidong Wang

The rise of large language models (LLMs) has transformed healthcare by offering clinical guidance, yet their direct deployment to patients poses safety risks due to limited domain expertise. To mitigate this, we propose repositioning LLMs as clinical assistants that collaborate with experienced physicians rather than interacting with patients directly. We co

Yifeng Zhu, Xianlin Zhao, Xutian Li, Yanzhen Zou

Repository summarization is a crucial research question in development and maintenance for software engineering. Existing repository summarization techniques primarily focus on summarizing code according to the directory tree, which is insufficient for tracing high-level features to the methods that collaboratively implement them. To address these limitation

K. Y. Zhang, X. X. Lu

The year 2025 marks the 40th anniversary of the discovery of halo nuclei and the 15th anniversary of the development of the deformed relativistic Hartree-Bogoliubov theory in continuum (DRHBc). In this work, we present the first DRHBc description of the proton halo phenomenon. The available experimental proton separation energies and empirical matter root-me

Sabine Hossenfelder

I present a simple argument for why a fundamental theory that unifies matter and gravity gives rise to what seems to be a collapse of the wavefunction. The resulting model is local, parameter-free and makes testable predictions.

Yeonseo Lee, Jungwook Mun, Hyosup Shin, Guebin Hwang

Real-world robotic systems frequently require diverse end-effectors for different tasks, however most existing grasp detection methods are optimized for a single gripper type, demanding retraining or optimization for each novel gripper configuration. This gripper-specific retraining paradigm is neither scalable nor practical. We propose XGrasp, a real-time g

Zixi Wang, Yushe Cao, Yubo Huang, Jinzhu Wei

In this paper, we propose a new method called Self-Training with Dynamic Weighting (STDW), which aims to enhance robustness in Gradual Domain Adaptation (GDA) by addressing the challenge of smooth knowledge migration from the source to the target domain. Traditional GDA methods mitigate domain shift through intermediate domains and self-training but often su

Longjie Guo, Chenjie Yuan, Mingyuan Zhong, Robert Wolfe

As LLM-based computer-use agents (CUAs) begin to autonomously interact with real-world interfaces, understanding their vulnerability to manipulative interface designs becomes increasingly critical. We introduce SusBench, an online benchmark for evaluating the susceptibility of CUAs to UI dark patterns, designs that aim to manipulate or deceive users into tak

Daisuke Ikegami, W. Hugh Woodin

We show that the Axiom of Real Determinacy $\mathsf{AD}_{\mathbb{R}}$ and the Axiom of Real Blackwell Determinacy $\mathsf{Bl}\text{-}\mathsf{AD}_{\mathbb{R}}$ are equivalent in $\mathsf{ZF}$+$\mathsf{DC}$. This answers the question of L\"{o}we [15, Question 53]. While we do not know whether they are equivalent in $\mathsf{ZF}$+$\mathsf{AC}_{\omega} (\mathbb

Kuldeep J. Purohit, Jitesh R. Bhatt, Subhendra Mohanty, Prashant K. Mehta

We perform a linear mode analysis of a uniformly distributed cloud of axion-like particles (ALPs) embedded in a magnetized intergalactic medium, in order to investigate the stability of axion stars under realistic astrophysical conditions. We find that when the frequency $\omega$ of transverse waves is much smaller than the collision frequency $\nu_c$ of the

Yiwei Liu, Yucheng Li, Xiao Li, Gong Cheng

Joint logical-numerical reasoning remains a major challenge for language models, yet existing datasets rely on fixed rule sets and offer limited control over task complexity, constraining their generalizability for evaluation and training. We present LogiNumSynth, a flexible natural language problem synthesizer that synthesizes tasks requiring proficiency in

Ji-Hoon Ha, Ibragim Alikhanov

Blazars, particularly Flat Spectrum Radio Quasars (FSRQs), are well-known for their ability to accelerate a substantial population of electrons and positrons, as inferred from multiwavelength radiation observations. Therefore, these astrophysical objects are promising candidates for studying high-energy electron--positron interactions, such as the production

Koji Kobashi

An issue concerning the Frenkel line of liquid CO2 is that its location has not been unequivocally determined. So far, reliable Frenkel lines were identified from velocity autocorrelation functions (VAFs) computed by Molecular Dynamic simulations; however VAFs cannot be directly verified experimentally. By contrast, the co-ordination numbers (CNs) can be mea

Yanning Hou, Ke Xu, Junfa Li, Yanran Ruan

Recently, the powerful generalization ability exhibited by foundation models has brought forth new solutions for zero-shot anomaly segmentation tasks. However, guiding these foundation models correctly to address downstream tasks remains a challenge. This paper proposes a novel two-stage framework, for zero-shot anomaly segmentation tasks in industrial anoma

Ganlin Yang, Tianyi Zhang, Haoran Hao, Weiyun Wang

While significant research has focused on developing embodied reasoning capabilities using Vision-Language Models (VLMs) or integrating advanced VLMs into Vision-Language-Action (VLA) models for end-to-end robot control, few studies directly address the critical gap between upstream VLM-based reasoning and downstream VLA policy learning. In this work, we tak

Hongxiang Li, Yaowei Li, Bin Lin, Yuwei Niu

Unified multimodal models integrate the reasoning capacity of large language models with both image understanding and generation, showing great promise for advanced multimodal intelligence. However, the community still lacks a rigorous reasoning-centric benchmark to systematically evaluate the alignment between understanding and generation, and their general

Alexander Plakhotnikov

The paper investigates the existence of a limit in the operator norm for a family of operators $T_z(H)= F(H-z)^{-1}F^*$ for $z$ tending to the real axis. The conditions for the $H$ operator and the rigging operator $F$ are established, under which the limit exists. Special attention is paid to the separation of cases when the limit point belongs and does not

X. Xie, A. Bergamaschi, M. Brückner, M. Carulla

The M\"ONCH hybrid pixel detector, with a 25 \textmu m pixel pitch and fast charge-integrating readout, has demonstrated subpixel resolution capabilities for X-ray imaging and deep learning-based electron localization in electron microscopy. Fully exploiting this potential requires extensive calibration to ensure both linearity and uniformity of the pixel re

Josefa Caballero, Łukasz Płociniczak, Kishin Sadarangani

We consider the initial-boundary value problem for a quasilinear time-fractional diffusion equation, and develop a fully discrete solver combining the parareal algorithm in time with a L1 finite-difference approximation of the Caputo derivative and a spectral Galerkin discretization in space. Our main contribution is the first rigorous convergence proof for

Lucas Wang, Jian Ge, Kevin Willis

Ca II absorbers, characterized by dusty and metal-rich environments, provide unique insights into the interstellar medium of galaxies. However, their rarity and weak absorption features have hindered comprehensive studies. In this work, we present a novel dual CNN approach to detect Ca II absorption systems, analyzing over 100,000 quasar spectra from the Slo

Clarissa R. Do Ó, Jaehan Bae, Quinn M. Konopacky, Jayke S. Nguyen

Direct imaging has revealed exoplanet systems hosting multiple wide-orbit Super-Jupiters, where planet-planet interactions can shape their long-term dynamical evolution. These strong perturbations may lead to orbital instability, raising questions about the long-term survival of such systems. Shortly after formation, planet-disk interactions can shepherd pla

Shasha Guo, Liang Pang, Xi Wang, Yanling Wang

Auxiliary lines are essential for solving complex geometric problems but remain challenging for large vision-language models (LVLMs). Recent attempts construct auxiliary lines via code-driven rendering, a strategy that relies on accurate and executable code generation to produce visual renderings of the auxiliary lines for subsequent reasoning. However, in c

Bingjie Tang, Iretiayo Akinola, Jie Xu, Bowen Wen

Simulation-based learning has enabled policies for precise, contact-rich tasks (e.g., robotic assembly) to reach high success rates (~80%) under high levels of observation noise and control error. Although such performance may be sufficient for research applications, it falls short of industry standards and makes policy chaining exceptionally brittle. A key

Pranav Ramesh, Arjun Roy, Deepak Ravikumar, Kaushik Roy

Designing adversarially robust models from a data-centric perspective requires understanding which input samples are most crucial for learning resilient features. While coreset selection provides a mechanism for efficient training on data subsets, current algorithms are designed for clean accuracy and fall short in preserving robustness. To address this, we

Runyang Feng, Hyung Jin Chang, Tze Ho Elden Tse, Boeun Kim

Modeling high-resolution spatiotemporal representations, including both global dynamic contexts (e.g., holistic human motion tendencies) and local motion details (e.g., high-frequency changes of keypoints), is essential for video-based human pose estimation (VHPE). Current state-of-the-art methods typically unify spatiotemporal learning within a single type

Ziyi Gao, Yike Xu, Jiahao Yuan, Baokun Wang

User representation modeling has become increasingly crucial for personalized applications, yet existing approaches struggle with generalizability across domains and sensitivity to noisy behavioral signals. We present InstructUE, an instruction-aware user embedding foundation model that leverages large language models (LLMs) to generate general and instructi

Yige Hong

Bounding the steady-state queue length of a multiserver queue is a central challenge in queueing theory. Even for the classical $GI/GI/n$ queue with homogeneous servers, obtaining a simple, accurate bound that holds across all parameters is highly non-trivial. A recent breakthrough by Li and Goldberg (2025) establishes the first universal bound of order $O(1

Tatsuru Kikuchi

This paper develops a unified framework for identifying spatial and temporal boundaries of treatment effects in difference-in-differences designs. Starting from fundamental fluid dynamics equations (Navier-Stokes), we derive conditions under which treatment effects decay exponentially in space and time, enabling researchers to calculate explicit boundaries b

Sanchit Sinha, Guangzhi Xiong, Aidong Zhang

Compositional reasoning remains a persistent weakness of modern vision language models (VLMs): they often falter when a task hinges on understanding how multiple objects, attributes, and relations interact within an image. Multiple research works have attempted to improve compositionality performance by creative tricks such as improving prompt structure, cha

Yi Wang, Yinfeng Yu, Fuchun Sun, Liejun Wang

Audio-Visual Embodied Navigation aims to enable agents to autonomously navigate to sound sources in unknown 3D environments using auditory cues. While current AVN methods excel on in-distribution sound sources, they exhibit poor cross-source generalization: navigation success rates plummet and search paths become excessively long when agents encounter unhear

Farzaneh Zirak, Farhana Choudhury, Renata Borovica-Gajic

Data prefetching--loading data into the cache before it is requested--is essential for reducing I/O overhead and improving database performance. While traditional prefetchers focus on sequential patterns, recent learning-based approaches, especially those leveraging data semantics, achieve higher accuracy for complex access patterns. However, these methods o

Nina Bonaventura, Jianwei Lyu, George H. Rieke, Andrew J. Bunker

In Paper I, we exploited the unsurpassed resolution and depth of JWST/NIRCam imagery to investigate the relationship between AGN and host-galaxy properties in the JWST era, finding a correlation between the level of spatial disturbance (as measured by shape asymmetry, $A_S$) and obscuration ($N_H$). Here in Paper II, we report an expansion of our X-ray and i

Jiamin Liang, Mingqiu Li, Yu Gao, Wei Ji

The observation of gravitational waves has opened a new window into the Universe through gravitational-wave astronomy. However, high-frequency gravitational waves remain undetected. In this work, we propose that spin systems can be employed to detect gravitational waves in this unexplored frequency regime. We derive the spin's response to gravitational waves

Ta-Chung Chi, Ting-Han Fan, Raffaele M. Ghigliazza, Domenico Giannone

We forecast the full conditional distribution of macroeconomic outcomes by systematically integrating three key principles: using high-dimensional data with appropriate regularization, adopting rigorous out-of-sample validation procedures, and incorporating nonlinearities. By exploiting the rich information embedded in a large set of macroeconomic and financ

Antonina Nepeivoda, Ilya Afanasyev

We introduce a string-interval abstract domain, where string intervals are characterized by systems of word equations (encoding lower bounds on string values) and word disequalities (encoding upper bounds). Building upon the lattice structure of string intervals, we define an abstract string object as a reduced product on a string property semilattice, deter

Jesse Atuhurra, Hidetaka Kamigaito, Taro Watanabe, Koichiro Yoshino

We introduce J-ORA, a novel multimodal dataset that bridges the gap in robot perception by providing detailed object attribute annotations within Japanese human-robot dialogue scenarios. J-ORA is designed to support three critical perception tasks, object identification, reference resolution, and next-action prediction, by leveraging a comprehensive template

Hai-Bo Yu

We show that core-collapsed self-interacting dark matter halos of mass $\sim 10^6\,{M_\odot}$, originally simulated to explain the dense perturber of the GD-1 stellar stream, also reproduce the structural properties inferred for the dense perturber detected in the strong lensing system JVAS B1938+666 from radio observations. Furthermore, these halos are suff

Kai Han, Siqi Ma, Chengxuan Qian, Jun Chen

Accurate segmentation of tumors and adjacent normal tissues in medical images is essential for surgical planning and tumor staging. Although foundation models generally perform well in segmentation tasks, they often struggle to focus on foreground areas in complex, low-contrast backgrounds, where some malignant tumors closely resemble normal organs, complica

Chenyu Zhang, Sharifa Alghowinem, Cynthia Breazeal

While recent studies have examined the leaning impact of large language model (LLM) in educational contexts, the affective dynamics of LLM-mediated tutoring remain insufficiently understood. This work introduces the first ensemble-LLM framework for large-scale affect sensing in tutoring dialogues, advancing the conversation on responsible pathways for integr

Haoran Liang, Yufa Zhou, Mohammad Talebi Kalaleh, Qipei Mei

We introduce $\textbf{MASSE}$, the first Multi-Agent System for Structural Engineering, effectively integrating large language model (LLM)-based agents with real-world engineering workflows. Structural engineering is a fundamental yet traditionally stagnant domain, with core workflows remaining largely unchanged for decades despite its substantial economic i

Takuma Fujiu, Sho Okazaki, Kohei Kaminishi, Yuji Nakata

In manufacturing systems, identifying the causes of failures is crucial for maintaining and improving production efficiency. In knowledge-based failure-cause inference, it is important that the knowledge base (1) explicitly structures knowledge about the target system and about failures, and (2) contains sufficiently long causal chains of failures. In this s

Z. L. Li, A. R. Sun, J. H. Xia, J. X. Wu

The electron-positron pair creation in a supercritical static asymmetric potential well, which is composed of a subcritical and a supercritical potential separated by a fixed distance, is investigated using computational quantum field theory. To explain the discrete peaks in the positron energy spectrum, an analytical formula for determining the positions of

Andrés Camilo Zúñiga-González, Anil Madhavapeddy, Ronita Bardhan

Nature access is increasingly recognised as a public health and equity imperative, yet cities lack standardised ways to measure who benefits from green infrastructure. We present the first national, building-level assessment of the 3-30-300 urban greening rule across England, integrating high-performance computing with open LiDAR and geospatial datasets. Our

Tieyuan Chen, Xiaodong Chen, Haoxing Chen, Zhenzhong Lan

We introduce Dynamic Nested Depth (DND), a novel method that improves performance for off-the-shelf LLMs by selecting critical tokens to reprocess in a nested depth manner. Specifically, at the end of the given transformer layer, DND identifies more critical tokens with a router and feeds them back for an extra round of processing, effectively ``reviewing" d

Ruihang Xu, Dewei Zhou, Fan Ma, Yi Yang

Multi-instance image generation (MIG) remains a significant challenge for modern diffusion models due to key limitations in achieving precise control over object layout and preserving the identity of multiple distinct subjects. To address these limitations, we introduce ContextGen, a novel Diffusion Transformer framework for multi-instance generation that is

Mahika Phutane, Hayoung Jung, Matthew Kim, Tanushree Mitra

Large language models (LLMs) are increasingly under scrutiny for perpetuating identity-based discrimination in high-stakes domains such as hiring, particularly against people with disabilities (PwD). However, existing research remains largely Western-centric, overlooking how intersecting forms of marginalization--such as gender and caste--shape experiences o

Jose M. Betancourt

I study dynamic network formation games in which agents meet stochastically and form links based on their valuation of the network. I show that these games can be represented in terms of the values agents assign to network sub-structures. Particularly, this characterizes potential games as those where all participants in a structure value it equally. When va

Ethan Abraham, Junghyun Yoon, Troy Van Voorhis, Martin Z. Bazant

The Marcus theory expression for the rate of non-adiabatic electron transfer is widely used across a range of physical conditions. Although Marcus theory defines the reorganization energy classically, here we show that the reorganization parameter appearing in the activation barrier for normal-region electron transfer is most generally a quantum mechanical o

Yongyi Zang, Jiarui Hai, Wanying Ge, Qiuqiang Kong

Music Source Restoration (MSR) extends source separation to realistic settings where signals undergo production effects (equalization, compression, reverb) and real-world degradations, with the goal of recovering the original unprocessed sources. Existing benchmarks cannot measure restoration fidelity: synthetic datasets use unprocessed stems but unrealistic

Wei-Chieh Huang, Henry Peng Zou, Yaozu Wu, Dongyuan Li

Deep research frameworks have shown promising capabilities in synthesizing comprehensive reports from web sources. While deep research possesses significant potential to address complex issues through planning and research cycles, existing frameworks are deficient in sufficient evaluation procedures and stage-specific protections. They typically treat evalua

Yuxin Cheng, Binxiao Huang, Taiqiang Wu, Wenyong Zhou

3D Gaussian inpainting, a critical technique for numerous applications in virtual reality and multimedia, has made significant progress with pretrained diffusion models. However, ensuring multi-view consistency, an essential requirement for high-quality inpainting, remains a key challenge. In this work, we present PAInpainter, a novel approach designed to ad

Uddalak Mukherjee, Sumit Som

Remotal and uniquely remotal sets play an important role in the area of farthest point problem as well as nearest point problem in a Banach space $X.$ In this study, we find some sufficient conditions for remotality and uniquely remotality of a bounded subset of a Banach space $X$ through $\alpha\beta$-statistical convergence.

Huanjin Yao, Ruifei Zhang, Jiaxing Huang, Jingyi Zhang

With the recent emergence of revolutionary autonomous agentic systems, research community is witnessing a significant shift from traditional static, passive, and domain-specific AI agents toward more dynamic, proactive, and generalizable agentic AI. Motivated by the growing interest in agentic AI and its potential trajectory toward AGI, we present a comprehe

Tianze Wang, Zhaoyu Chen, Jian Du, Yingtai Xiao

Text data has become extremely valuable on large language models (LLMs) and even lead to general artificial intelligence (AGI). A lot of high-quality text in the real world is private and cannot be freely used due to privacy concerns. Therefore, differentially private (DP) synthetic text generation has been proposed, aiming to produce high-utility synthetic

Shivanshu Kumar, Gopalakrishnan Srinivasan

While the transformer architecture has achieved state-of-the-art performance on natural language processing tasks, these models impose substantial memory and computational overhead. Recent research has identified significant architectural redundancies within these models, particularly in the attention sub-layers in the top layers, presenting opportunities fo

Yixiong Gao, Florian Jaehn, Minming Li, Wenhao Ma

During loading and unloading steps, energy is consumed when cranes lift containers, while energy is often wasted when cranes drop containers. By optimizing the scheduling of cranes, it is possible to reduce energy consumption, thereby lowering operational costs and environmental impacts. In this paper, we introduce a single-crane scheduling problem with ener

Hyeseon An, Shinwoo Park, Suyeon Woo, Yo-Sub Han

The promise of LLM watermarking rests on a core assumption that a specific watermark proves authorship by a specific model. We demonstrate that this assumption is dangerously flawed. We introduce the threat of watermark spoofing, a sophisticated attack that allows a malicious model to generate text containing the authentic-looking watermark of a trusted, vic

Xiaoyu Ma, Ding Ding, Hao Chen

Multimodal video understanding plays a crucial role in tasks such as action recognition and emotion classification by combining information from different modalities. However, multimodal models are prone to overfitting strong modalities, which can dominate learning and suppress the contributions of weaker ones. To address this challenge, we first propose Mul

Menghan Yi, Yanlin Tang, Huixia Judy Wang

In real-world applications, the limited availability of labeled outcomes presents significant challenges for statistical inference due to high collection costs, technical barriers, and other constraints. In this work, we propose a method to construct efficient conformal prediction sets for new target outcomes by leveraging a source distribution that is disti

Jingyi Wang, Nai-Yuan Chiang, Tucker Hartland, J. Luc Peterson

Recently, multi-fidelity Bayesian optimization (MFBO) has been successfully applied to many engineering design optimization problems, where the cost of high-fidelity simulations and experiments can be prohibitive. However, challenges remain for constrained optimization problems using the MFBO framework, particularly in efficiently identifying the feasible re

Wenbing Jiang, Xuankai Xu, Jiazhen Pan, Hancong Sun

Lamb wave resonators (LWRs) operating at ultralow temperatures serve as promising acoustic platforms for implementing microwave-optical transduction and radio frequency (RF) front-ends in aerospace communications because of the exceptional electromechanical coupling (k^2) and frequency scalability. However, the properties of LWRs at cryogenic temperatures ha

Di Zhang

Self-supervised learning (SSL) has achieved remarkable success by learning meaningful representations without labeled data. However, a unified theoretical framework for understanding and comparing the efficiency of different SSL paradigms remains elusive. In this paper, we introduce a novel information-geometric framework to quantify representation efficienc

Qizhi Guo, Siyuan Yang, Junning Lyu, Jianjun Sun

Accurate and robust heading estimation is crucial for unmanned aerial vehicles (UAVs) when conducting indoor inspection tasks. However, the cluttered nature of indoor environments often introduces severe magnetic disturbances, which can significantly degrade heading accuracy. To address this challenge, this paper presents an Adaptive MARG-Only Heading (AMO-H

Bohao Wang, Jiawei Chen, Feng Liu, Changwang Zhang

Large language models (LLMs), owing to their extensive open-domain knowledge and semantic reasoning capabilities, have been increasingly integrated into recommender systems (RS). However, a substantial gap remains between the pre-training objectives of LLMs and the specific requirements of recommendation tasks. To address this gap, supervised fine-tuning (SF

Taiqiang Wu, Runming Yang, Tao Liu, Jiahao Wang

Model merging, typically on Instruct and Thinking models, has shown remarkable performance for efficient reasoning. In this paper, we systematically revisit the simplest merging method that interpolates two weights directly. Particularly, we observe that model interpolation follows a three-stage evolutionary paradigm with distinct behaviors on the reasoning

Wentao Wang, Heqing Zou, Tianze Luo, Rui Huang

Recent progress in Multimodal Large Language Models (MLLMs) has demonstrated strong semantic understanding capabilities, but struggles to perform precise spatio-temporal understanding. Existing spatio-temporal methods primarily focus on the video itself, while overlooking the physical information within the video, such as multi-object layouts and motion. Suc

Mingtong Dai, Lingbo Liu, Yongjie Bai, Yang Liu

Vision-Language-Action (VLA) models have become a prominent paradigm for embodied intelligence, yet further performance improvements typically rely on scaling up training data and model size -- an approach that is prohibitively expensive for robotics and fundamentally limited by data collection costs. We address this limitation with $\mathbf{RoVer}$, an embo