2026-10论文中文摘要
表示动力学揭示多模态大语言模型中视觉标记剪枝的语义显著性与相似性
Representation Dynamics Reveal Semantic Saliency and Similarity for Visual Token Pruning in MLLMs
arXiv 2609.36916 · 2026-10-01RoXDrive：基于动作忠实回滚的端到端自动驾驶闭环强化学习
RoXDrive: Closed-Loop Reinforcement Learning for End-to-End Autonomous Driving via Action-Faithful Rollouts
arXiv 2609.36851 · 2026-10-01延迟微分方程的偏移线性正则变换分析
Analysis of Delay Differential Equations Using the Offset Linear Canonical Transform
arXiv 2609.36792 · 2026-10-01编辑器只有只读权限：扩散语言模型中的正确性信号
A Heckler in the Hidden State: Correctness Signals in Diffusion Language Models
arXiv 2609.36783 · 2026-10-01QuantMLA: 面向低比特 MLA KV 缓存的功能对齐双路径量化
QuantMLA: Function-Aligned Dual-Path Quantization for Low-Bit MLA KV Caching
arXiv 2609.36760 · 2026-10-01紧球面$5$-设计的算术不存在性结果
Arithmetic Nonexistence Conditions for Tight Spherical $5$-Designs
arXiv 2609.36725 · 2026-10-01裂变与聚变能源系统中氚释放的综合评估及其环境影响
Integrated assessment of tritium releases from fission and fusion energy systems and their environmental implications
arXiv 2609.36713 · 2026-10-01人类启发的任务维度引导探索：高维空间中的高效学习
Human-inspired, Task-Dimension-Guided Exploration for Efficient Learning in High Dimensions
arXiv 2609.36672 · 2026-10-01玄武质熔体粘度的异常压力依赖性及其在软流圈熔体聚集中的作用
Anomalous pressure-dependent viscosity of basaltic melts and its role in asthenosphere melt accumulation
arXiv 2609.36623 · 2026-10-01超越可读性：统一视频生成中视觉文本渲染与就地编辑的基准评测
Beyond Legibility: Benchmarking Visual Text Rendering and In-Place Editing in Unified Video Generation
arXiv 2609.36598 · 2026-10-01Text2Sim：基于蒸馏专业知识的智能体物理仿真生成
Text2Sim: Agentic Physics-Based Simulation Generation with Distilled Expertise
arXiv 2609.36593 · 2026-10-01用于MIMO设备到设备干扰网络的代理增强分数规划
Surrogate-Enhanced Fractional Programming for MIMO Device-to-Device Interference Networks
arXiv 2609.36586 · 2026-10-01PDE-OBS：跨观测模式的受控评估
PDE-OBS: Controlled Evaluation Across Observation Patterns
arXiv 2609.36521 · 2026-10-01恐音症触发声音抑制
Trigger Sound Suppression for Misophonia
arXiv 2609.36351 · 2026-10-01ATLAS：潜在结构的对齐传输用于可靠的世界模型规划
ATLAS: Aligned Transport of Latent Structure for Reliable World Model Planning
arXiv 2609.36333 · 2026-10-01PILLAR：用于增强检索的私有倒排索引词法查找
PILLAR: Private Inverted-Index Lexical Lookup for Augmented Retrieval
arXiv 2609.36326 · 2026-10-01量化分布式酉耦合簇拟设中的隐形传态开销
Quantifying Teleportation Overhead in Distributed Unitary Coupled-Cluster Ansätze
arXiv 2609.36266 · 2026-10-01稀疏视角可解释的三维动物行为表征用于神经编码与解码
Sparse-View Interpretable 3D Animal Behavior Representations for Neural Encoding and Decoding
arXiv 2609.36217 · 2026-10-01PHASE：一种面向颅内脑电图的生理引导分层基础模型
PHASE: A Physiology-Guided Hierarchical Foundation Model for Intracranial EEG
arXiv 2609.36087 · 2026-10-01Argus：生成式人工智能时代的学术诚信
Argus: Academic Integrity in the Era of Generative AI
arXiv 2609.36073 · 2026-10-01非线性无序哈密顿系统中的相干能量级联
Coherent energy cascades in nonlinear disordered Hamiltonians
arXiv 2609.36009 · 2026-10-01多项式系统的伪解与$\mbox{AC}^0[p]$-Frege系统的下界问题
Pseudo-solutions of polynomial systems and the lower bound problem for $\mbox{AC}^0[p]$-Frege systems
arXiv 2609.35927 · 2026-10-01关于Talagrand型
On Talagrand Type
arXiv 2609.35899 · 2026-10-01固定价格双边交易的确切福利保证
The Exact Welfare Guarantee of Fixed-Price Bilateral Trade
arXiv 2609.35888 · 2026-10-01Transformer 权重的介观视角：行与列尺度场
A Mesoscopic View of Transformer Weights Through Row and Column Scale Fields
arXiv 2609.35852 · 2026-10-01元学习加速探测器设计优化
Meta-learning accelerates detector design optimization
arXiv 2609.35827 · 2026-10-01Sage：带语义校正的形式化
Sage: Formalization with Semantic Correction
arXiv 2609.35790 · 2026-10-01Muon Bulb：灯泡内的宇宙射线探测器
Muon Bulb: A Cosmic-Ray Detector Inside a Light Bulb
arXiv 2609.35784 · 2026-10-01化学计量材料中共线磁结构目录
Catalogue of Collinear Magnetic Structures in Stoichiometric Materials
arXiv 2609.35712 · 2026-10-01K矮星黑子搜寻者：如何利用紫外-红外全波段凌星光谱直接表征黑子和谱斑性质
K-dwarf starspot hunters: how to directly characterize spot and faculae properties with UV-IR panchromatic transit spectra
arXiv 2609.35676 · 2026-10-01从嘈杂到层级：一个评估AI意识的原则性框架
From cacophony to hierarchy: a principled framework for assessing AI consciousness
arXiv 2609.35618 · 2026-10-01SEABench：自我进化智能体中的内生错位基准测试
SEABench: Benchmarking Endogenous Misalignment In Self-Evolving Agents
arXiv 2609.35596 · 2026-10-01RSI-Master：结构化实验以引导自主模型改进
RSI-Master: Structuring Experiments to Guide Autonomous Model Improvement
arXiv 2609.35561 · 2026-10-01从分数到样本：自回归视频生成的弹性强制
From Scores to Samples: Elastic Forcing for Autoregressive Video Generation
arXiv 2609.35491 · 2026-10-01AHMAD：用于关键点检测的自适应混合多任务视觉学习与辅助蒸馏
AHMAD: Adaptive Hybrid Multi-task Vision Learning with Assisted Distillation for Keypoint Detection
arXiv 2609.35490 · 2026-10-01预训练图像编码器中的表征风险
Representation Risk in Pretrained Image Encoders
arXiv 2609.35470 · 2026-10-01Uni-VLaT：用于人形机器人全身移动操作的全身体触觉自适应VLA策略
Uni-VLaT: Whole-Body Tactile Adaptation of VLA Policies for Humanoid Loco-Manipulation
arXiv 2609.35450 · 2026-10-01稳定Higgs丛的点态能量单调性
Pointwise energy monotonicity for stable Higgs bundles
arXiv 2609.35278 · 2026-10-01Szemerédi-Trotter估计在有限域中的及其应用
The Szemerédi-Trotter Estimate in Finite Field with its Applications
arXiv 2609.35190 · 2026-10-01介电常数近零纳米球中的非线性量子光学：宏观量子电动力学方法
Nonlinear Quantum Optics in Epsilon-Near-Zero Nanospheres: a Macroscopic Quantum Electrodynamics Approach
arXiv 2609.35146 · 2026-10-01双层石墨烯中的门反射测量电容传感
Capacitance sensing in bilayer graphene with gate reflectometry
arXiv 2609.35125 · 2026-10-01学习高风险高精度运动控制
Learning High-Risk High-Precision Motion Control
arXiv 2609.34851 · 2026-10-01D$^2$-VLA：面向长时动态操作的双记忆双频视觉-语言-动作模型
D$^2$-VLA: Dual-Memory Dual-Frequency Vision-Language-Action Model For Long Dynamic Manipulation
arXiv 2609.34792 · 2026-10-01面向演化的属性图模式的变换方法
Transformations for Evolving Property Graph Schemas
arXiv 2609.34789 · 2026-10-01基于网格基础模型监督的毫米波雷达隐私保护全身网格重建
Privacy-Preserving Full-Body Meshing from mmWave Radar via Mesh Foundation Model Supervision
arXiv 2609.34768 · 2026-10-01目标试验模拟中的时间划分
Partitioning Time in Target Trial Emulation
arXiv 2609.34576 · 2026-10-01调和条件流匹配中的频谱演化以用于文本到语音合成
Harmonizing Spectral Evolution in Conditional Flow Matching for TTS
arXiv 2609.34431 · 2026-10-01编码智能体记忆后训练：通过强化学习解锁预训练文件操作的长时程任务记忆潜力
Coding Agent Memory Post-training: Unlocking the Memory Potential of Pre-trained File Operations for Long-Horizon Tasks via Reinforcement Learning
arXiv 2609.34422 · 2026-10-01全向高斯泼溅的率失真自适应基元选择
Rate-Distortion Adaptive Primitive Selection for Omnidirectional Gaussian Splatting
arXiv 2609.34367 · 2026-10-01MiCo：通过语义擦除建模实现互信息覆盖优化，用于高效 MLLM 推理
MiCo: Mutual Information Coverage Optimization through Semantic Erasure Modeling for Efficient MLLM Inference
arXiv 2609.34330 · 2026-10-01
