2026-09论文中文摘要
ArGuard 共享任务：阿拉伯语梗图与 LLM 提示词中的有害内容检测
ArGuard Shared Task: Harmful Content Detection in Arabic Memes and LLM Prompts
arXiv 2609.29349 · 2026-09-28EgoSpeedUp：将人类操作节奏迁移至机器人策略
EgoSpeedUp: Transferring Human Manipulation Tempo to Robot Policies
arXiv 2609.29310 · 2026-09-28涌现对称商及其弱展开的信息几何
Information geometry of emergent symmetry quotients and their weak unfoldings
arXiv 2609.29260 · 2026-09-28燃料分层长度尺度对热扩散不稳定贫燃氢火焰的影响
Effect of fuel stratification length scale on thermodiffusively unstable lean hydrogen flames
arXiv 2609.29201 · 2026-09-28编排AI辅助代码修复：大型工业代码库中的社会技术瓶颈
Orchestrating AI-Assisted Code Remediation: Socio-Technical Bottlenecks in a Large Industrial Repository
arXiv 2609.29172 · 2026-09-28可逆常数Q变换的精确分解与快速计算
Exact Factorisation and Fast Computation of Invertible Constant-Q Transforms
arXiv 2609.29119 · 2026-09-28一种用于连续体机械臂基于强化学习抓取的支持增强型颗粒阻塞夹爪
A Support-Enhanced Granular-Jamming Gripper for RL-based Grasping with Continuum Manipulators
arXiv 2609.29093 · 2026-09-28M-plicits：基于嵌套多尺度残差的神经隐式曲面
M-plicits: Neural Implicit Surfaces via Nested Multiscale Residuals
arXiv 2609.28684 · 2026-09-28基于矢量势的束流电流传感器：在 Mu2e 中的应用及皮秒级束团子结构
Vector Potential-based Beam Current Sensors: Applications for Mu2e, and Picosecond Bunch Substructure
arXiv 2609.28663 · 2026-09-28谱几何：旧问题与新答案
Spectral geometry: old questions and new answers
arXiv 2609.28602 · 2026-09-28随机惯性Krasnosel'skii-Mann迭代实现近最优样本复杂度
Stochastic Inertial Krasnosel'skii-Mann Iteration Achieves Near-Optimal Sample Complexity
arXiv 2609.28543 · 2026-09-28拆除脚手架的代价：麦克斯韦关于安培的论述，多年后竟被施加于他自身
The Price of Removing the Scaffolding: Maxwell wrote something about Ampère that years later would be done to him
arXiv 2609.28525 · 2026-09-28科学中的AI：早期洞察
AI in Science: Early Insights
arXiv 2609.28504 · 2026-09-28转导学习的更紧界及其应用
Even Sharper Bounds for Transductive Learning and Its Applications
arXiv 2609.28459 · 2026-09-28LightMIS：无需逐阶段解码器的超轻量级医学图像分割
LightMIS: Ultra-Lightweight Medical Image Segmentation Without a Stage-Wise Decoder
arXiv 2609.28327 · 2026-09-28支持编译的特征折叠：在表格基础模型中用更低内存获得更多证据
Support-Compiled Feature Folding: More Evidence at Lower Memory Across Tabular Foundation Models
arXiv 2609.28208 · 2026-09-28纵向血糖表征的可迁移证据重建
Transferable Evidence Reconstruction for Longitudinal Glucose Representations
arXiv 2609.28199 · 2026-09-28具有容许例外李对称性的含对数非线性项的多维抛物方程
Multidimensional parabolic equations with a logarithmic nonlinearity admitting exceptional Lie symmetries
arXiv 2609.27810 · 2026-09-28NS-ATTENTION：视觉Transformer中注意力输出的牛顿-舒尔茨变换
NS-ATTENTION: Newton-Schulz Transformations of Attention Outputs in Vision Transformers
arXiv 2609.27735 · 2026-09-28反驳QAOA固定角度猜想
Refuting the QAOA fixed-angle conjecture
arXiv 2609.27703 · 2026-09-28表格基础模型在上下文中计算什么？通过注意力门控更新的原位表示精炼
What Do Tabular Foundation Models Compute In Context? In-Situ Representation Refinement through Attention-Gated Updates
arXiv 2609.27679 · 2026-09-28Track2Art：从二维点跟踪器恢复以运动为中心的铰接物体模型
Track2Art: Articulated Object Model Recovery with Visual-Geometric Track Representations
arXiv 2609.27675 · 2026-09-28NV-Reason-CT：用于CT分析的3D视觉语言模型
NV-Reason-CT: 3D Visual Language Model for CT Analysis
arXiv 2609.27511 · 2026-09-28EmbodiedSWE：面向长时程灵巧机器人的编码智能体
EmbodiedSWE: Coding Agents for Long Horizon Dexterous Robotics
arXiv 2609.27308 · 2026-09-28超越重叠：估计基准暴露的因果效应
LeakScale: Estimating the Causal Effect of Benchmark Exposure
arXiv 2609.27176 · 2026-09-28黎曼-洛伦兹延拓的拓扑障碍
Topological Obstructions to Riemannian-Lorentzian Continuation
arXiv 2609.27136 · 2026-09-28基于子任务分解的强化学习
Reinforcement Learning with Decomposed Subtasks
arXiv 2609.27035 · 2026-09-28虫洞喉部有限宽度涡旋的标量晕与锥形匹配
Scalar halos and conical matching for finite-width vortices at a wormhole throat
arXiv 2609.26901 · 2026-09-28保持稳定性的半群生成元、谱间隙与经典基态计算
Generators of stability-preserving semigroups, spectral gaps, and classical ground-state computation
arXiv 2609.26802 · 2026-09-28校准诱饵：对抗性指纹识别下的防御性信息设计
Calibrated Bait: Defensive Information Design under Adversarial Fingerprinting
arXiv 2609.26797 · 2026-09-28面向多电网潮流的层次化图神经网络：跨运行场景的泛化
Hierarchical GNNs for power flow: letting physics shape the hierarchy
arXiv 2609.26603 · 2026-09-28超越最终任务成功率：如何审计机器人中的视觉经验检索
Beyond End-Task Success: How to Audit Visual Experience Retrieval in Robotics
arXiv 2609.26567 · 2026-09-28二面体轨道产生的无理Seshadri常数
Irrational Seshadri constants from dihedral orbits
arXiv 2609.26521 · 2026-09-28几何感知的双曲残差量化
Geometry-Aware Hyperbolic Residual-Quantized Variational Autoencoders
arXiv 2609.26342 · 2026-09-28不可控变量：视觉语言模型的拒答行为对图像存在性的响应方式无法用风险解释
The Uncontrolled Variable: Vision-Language Refusal Is Conditioned on the Image-Attachment Interface, and Not Robust to Irrelevant Image Properties
arXiv 2609.26174 · 2026-09-28动力学、热学与基态多参数量子计量学：基本极限及其在磁强测量中的可达性
Dynamical, thermal, and ground-state multiparameter quantum metrology: Fundamental limits and their attainability in magnetometry
arXiv 2609.25977 · 2026-09-28完全图关联族的精确二阶Zarankiewicz数
Exact Second-Order Zarankiewicz Numbers for Complete-Graph Incidence Families
arXiv 2609.25974 · 2026-09-28基于后选择机制的随机量子电池
Post-Selection-Based Stochastic Quantum Battery
arXiv 2609.25699 · 2026-09-28可扩展最小体积单纯形估计及其非渐近分析
Scalable Minimum-Volume Simplex Estimation with Non-asymptotic Analysis
arXiv 2609.25576 · 2026-09-28希尔采样用于测试时扩展：一种比重复采样、进化和训练更简单且更优的替代方案
Hill Sampling for Test-Time Scaling: A Simple and Better Alternative to Repeated Sampling, Evolution, and Training
arXiv 2609.25510 · 2026-09-28MIND the Gap：一种具有可调空间尺度的地理隐式神经表示
MIND the Gap: A Geographic Implicit Neural Representation with Adjustable Spatial Scale
arXiv 2609.25454 · 2026-09-28你已经看得够多了：面向机器的质量约束图像编码
You've Seen Enough: Quality-Constrained Image Coding for Machines
arXiv 2609.25108 · 2026-09-28通过多模态可穿戴传感检测自闭症青少年行为升级前的躁动
Detecting Agitation Before Behavioral Escalation in Autistic Youth Through Multimodal Wearable Sensing
arXiv 2609.24791 · 2026-09-28高阶逻辑自动定理证明器证明的形式验证
Formal Verification of Proofs from Automated Theorem Provers for Higher-Order Logic
arXiv 2609.24594 · 2026-09-28推动VLTI上从Y波段到L波段的高角分辨率和高对比度观测：Asgard仪器套件的集成状态与计划
Pushing high angular resolution and high contrast observations on the VLTI from Y to L band with the Asgard instrumental suite: integration status and plans
arXiv 2609.24567 · 2026-09-28利用光子芯片上的核零干涉与主动相位控制抑制恒星光芒
Silencing the stellar glare with kernel-nulling and active phase control on photonic chips
arXiv 2609.24540 · 2026-09-28Hopf纹理中由Navier–Stokes爆破类比驱动的折叠级联起始
Fold Catastrophes in a Hopf Texture Driven by a Navier--Stokes Blow-Up Analog
arXiv 2609.24490 · 2026-09-28提升的贝尔曼线性规划用于离线强化学习
Lifted Bellman Linear Programming for Offline Reinforcement Learning
arXiv 2609.24489 · 2026-09-28弦荷密度/比特线对应在II型超引力中的T-对偶性
T-duality of string charge density/bit threads correspondence in type II supergravity
arXiv 2609.24366 · 2026-09-28轴向热膨胀：基于温度相关力常数的自由能最小化方法在phonopy中的实现——技术报告，以$\alpha$-Ti和$\omega$-Ti为例
Axial thermal expansion from free-energy minimization with temperature-dependent force constants in phonopy: a technical report, with $α$- and $ω$-Ti as the worked example
arXiv 2609.24336 · 2026-09-28
