2026-09论文中文摘要
先验驱动的三维高斯泼溅增强：法线与深度正则化
Prior-Driven Enhancements in 3D Gaussian Splatting: Normals and Depths Regularization
arXiv 2609.36969 · 2026-09-30TaskBridge：通过虚拟任务桥接无监督表格异常检测与上下文学习
TaskBridge: Bridging Unsupervised Tabular Anomaly Detection and In-Context Learning via Virtual Tasks
arXiv 2609.36968 · 2026-09-30超越令牌重要性：保留空间支架以实现高效的视觉-语言-动作推理
Beyond Token Importance: Preserving Spatial Scaffolds for Efficient Vision-Language-Action Inference
arXiv 2609.36967 · 2026-09-30Chinese-Jev：将 System One 模型引入中文任务
Chinese-Jev: Bringing System One Model to Chinese-Language Tasks
arXiv 2609.36965 · 2026-09-30截断可调等离子体通道中的电子再相位
Electron Rephasing in a Truncated Tunable Plasma Channel
arXiv 2609.36964 · 2026-09-30融合超度量Gromov-Wasserstein距离用于多阶段随机规划中的场景树生成
Fused ultrametric Gromov-Wasserstein for Scenario Tree generation in Multistage Stochastic Programming
arXiv 2609.36963 · 2026-09-30深度学习GW准粒子哈密顿量用于大规模多体激发态电子结构
Deep Learning GW Quasiparticle Hamiltonians for Many-Body Excited-State Electronic Structure at Scale
arXiv 2609.36962 · 2026-09-30三星系统中的LISA双白矮星
LISA Double White Dwarfs in Triple Systems
arXiv 2609.36961 · 2026-09-30Cobalt：利用专家共激活实现高效的分布式MoE训练
Cobalt: Leveraging Expert Co-activation for Efficient Distributed MoE Training
arXiv 2609.36959 · 2026-09-30VStress：面向重复验证器的相关性感知审计与自适应预算分配
VStress: Correlation-Aware Auditing and Adaptive Budget Allocation for Repeated Verifiers
arXiv 2609.36958 · 2026-09-30超越可读性：评估任务信息可恢复性
Beyond Readability: Evaluating Task Information Recoverability
arXiv 2609.36957 · 2026-09-30黑盒大语言模型的受控解码攻击
Controlled Decoding Attacks on Black-Box LLMs
arXiv 2609.36956 · 2026-09-30信息欧姆定律：由斯格明子麦克斯韦妖揭示
Ohm's law for information revealed by a skyrmion Maxwell's demon
arXiv 2609.36955 · 2026-09-30Purlin：将编排与集合通信的数据路径分离
Purlin: Separating Orchestration from the Datapath of Collectives
arXiv 2609.36954 · 2026-09-30冷却采样器，而非学习者：采样温度移动重要性校正 GRPO 的陈旧性悬崖
Cool the Sampler, Not the Learner: Sampling Temperature Moves the Staleness Cliff of Importance-Corrected GRPO
arXiv 2609.36953 · 2026-09-30ER-JEPA：经验回放改进语言模型中的联合嵌入预测学习
ER-JEPA: Experience Replay Improves Joint-Embedding Predictive Learning in Language Models
arXiv 2609.36952 · 2026-09-30解释与评估动态码率语音编解码器边界
Interpreting and Evaluating Dynamic-Rate Speech Codec Boundaries
arXiv 2609.36951 · 2026-09-30可扩展扩散SBI用于模拟器误设下的组合推断
Scalable Diffusion SBI for Compositional Inference under Simulator Misspecification
arXiv 2609.36950 · 2026-09-30来自莱曼-$\alpha$ 森林的弱引力透镜测量
Weak Lensing Measurements from the Lyman-$α$ Forest
arXiv 2609.36949 · 2026-09-30利用切向次微分研究双层规划的必要最优性条件
Necessary Optimality Conditions for Bilevel Programming using Tangential Subdifferentials
arXiv 2609.36948 · 2026-09-30硬约束与软约束物理信息分子机器学习与Clapeyron方程的系统比较
A systematic comparison of hard- and soft-constrained physics-informed molecular machine learning with the Clapeyron equation
arXiv 2609.36947 · 2026-09-30优化VLP对齐的多模态意图表示与正确视觉实例化用于零样本组合图像检索
Optimizing VLP-aligned Multimodal Intent Representation with Correct Visual Instantiation for Zero-Shot Composed Image Retrieval
arXiv 2609.36946 · 2026-09-30基于自生成与奖励加权数据的微调：学习动态、收敛速率与离策略性的优势
Fine-Tuning on Self-Generated and Reward-Weighted Data: Learning Dynamics, Convergence Rates, and Benefits of Off-Policyness
arXiv 2609.36945 · 2026-09-30双通道鲁棒群相对策略优化：基于优势与序列权重估计
Dual-Channel Robust Group-Relative Policy Optimization via Advantage and Sequence-Weight Estimation
arXiv 2609.36944 · 2026-09-30宇宙学额外维度可以模拟暗物质 II. 宇宙学结构的增长
Cosmological extra dimensions can mimic dark matter II. Growth of cosmological structures
arXiv 2609.36943 · 2026-09-30针对黑盒大语言模型的实际秘密提取
Practical Secrets Extraction against Black-box LLMs
arXiv 2609.36941 · 2026-09-30DispFlow-GS：基于运动解耦的位移流监督用于单目可变形3D高斯泼溅
DispFlow-GS: Displacement Flow Supervision with Motion Disentangling for Monocular Deformable 3D Gaussian Splatting
arXiv 2609.36940 · 2026-09-30SCA：GUI代理强化学习的空间信用分配
SCA: Spatial Credit Assignment for Reinforcement Learning of GUI Agents
arXiv 2609.36939 · 2026-09-30基于SSD稀疏KV存储的高效智能体LLM服务
Efficient Agentic LLM Serving over SSD-based Sparse KV Storage
arXiv 2609.36938 · 2026-09-30WeLike2Party! 多人物图像动画中的上下文运动迁移
WeLike2Party! In-Context Motion Transfer for Multi-Human Image Animation
arXiv 2609.36937 · 2026-09-30空穴掺杂安德森自旋绝缘体中的遍历性
Ergodicity in a hole-dopped Anderson spin insulator
arXiv 2609.36936 · 2026-09-30VLALight：面向交通信号控制的视觉-语言-动作模型
VLALight: A Vision-Language-Action Model for Traffic Signal Control
arXiv 2609.36934 · 2026-09-30味分辨四费米子探针在FCC-ee $Z$极点处的研究
Flavor-resolved four-fermion probes at the FCC-ee $Z$ pole
arXiv 2609.36933 · 2026-09-30为模型标注日期：系统提示中的隐藏日期影响大语言模型评估
Dating the Model: Hidden Dates in System Prompts Affect LLM Evaluation
arXiv 2609.36931 · 2026-09-30NuPaD：一种促进亚原子物理学深度学习的生成式人工智能框架
NuPaD: A Generative AI Framework for Fostering Deep Learning in Subatomic Physics
arXiv 2609.36930 · 2026-09-30SFE-VGGT：基于事件的单目深度估计的无源VGGT蒸馏
SFE-VGGT: Source-Free VGGT Distillation for Event-Based Monocular Depth Estimation
arXiv 2609.36929 · 2026-09-30ComManip：将操作策略过拟合到舒适区域
ComManip: Overfitting Manipulation Policies to Comfortable Regions
arXiv 2609.36928 · 2026-09-30神经符号计算机使用：学习可重用策略以实现可靠高效的执行
Neuro-Symbolic Computer Use: Learning Reusable Policies for Reliable and Efficient Execution
arXiv 2609.36927 · 2026-09-30状态传输路由用于多时域光伏预测中的短期适应
State Transport Routing for Short-horizon Adaptation in Multi-horizon Photovoltaic Forecasting
arXiv 2609.36926 · 2026-09-30自旋-2 $QQ\bar Q\bar s$ 四夸克态的QCD求和规则研究
Spin-2 $QQ\bar Q\bar s$ Tetraquarks in QCD Sum Rules
arXiv 2609.36925 · 2026-09-30Track-and-Complete：从单个失败的人类视频中学习人形机器人技能
Track-and-Complete: Learning Humanoid Skills from a Single Failed Human Video
arXiv 2609.36924 · 2026-09-30PrecogUI：通过预认知模拟与经验检索实现主动式GUI智能体
PrecogUI: Proactive GUI Agents via Pre-cognitive Simulation and Experience Retrieval
arXiv 2609.36923 · 2026-09-30层的平移辛模叠的退化
Degenerations of shifted symplectic moduli stacks of sheaves
arXiv 2609.36922 · 2026-09-30当能力无法组合：诊断大型音频-语言模型中的组合性差距
When Capabilities Fail to Compose: Diagnosing the Compositionality Gap in Large Audio-Language Models
arXiv 2609.36921 · 2026-09-30伊比利亚语言自动语音识别工具基准评测
Benchmarking Automatic Speech Recognition Tools for Iberian Languages
arXiv 2609.36920 · 2026-09-30调谐原子层沉积合成MgZnO多层膜的二次电子产额和电子导电性
Tuning the secondary electron yield and electronic conductivity of MgZnO multilayers synthesized by Atomic layer deposition
arXiv 2609.36919 · 2026-09-30Seg3DParts：基于分割的可控部件级三维生成
Seg3DParts: Segmentation-Grounded Controllable Part-Level 3D Generation
arXiv 2609.36918 · 2026-09-30多模光纤中的横向法拉第效应
Transverse Faraday Effect in Multimode Fibres
arXiv 2609.36917 · 2026-09-30AeroManip-VLA：基于强化学习生成演示的可扩展空中操纵视觉-语言-动作学习
AeroManip-VLA: Scalable Vision-Language-Action Learning for Aerial Manipulation with RL-Generated Demonstrations
arXiv 2609.36915 · 2026-09-30语言模型能否学会预测股票价格
Can Language Models Learn to Forecast Stock Prices
arXiv 2609.36914 · 2026-09-30
