2026-10论文中文摘要
OnlineQAT：面向超低位大型语言模型的在线策略蒸馏
OnlineQAT: On-Policy Distillation for Ultra-Low-Bit Large Language Models
arXiv 2610.09346 · 2026-10-08使用离位$^3$He中子自旋过滤器和二维探测器的极化中子衍射用于非共线不相称磁序研究
Polarized neutron diffraction with ex-situ $^3$He neutron spin filter and two-dimensional detector for noncollinear incommensurate magnetic order
arXiv 2610.09345 · 2026-10-08图像编辑器是否遵循深度相关的模糊与光圈响应？一项基于渲染真值的试点审计
Do Image Editors Follow Depth-Dependent Blur and Aperture Response? A Rendered-Ground-Truth Pilot Audit
arXiv 2610.09344 · 2026-10-08TileSkipper：面向3D高斯泼溅的区域自适应瓦片剪枝
TileSkipper: Region-Adaptive Tile Pruning for 3D Gaussian Splatting
arXiv 2610.09343 · 2026-10-08共享低秩基分解用于无数据混合专家模型压缩
Shared Low-rank Basis Factorization for Data-free Mixture-of-Experts Compression
arXiv 2610.09342 · 2026-10-08利用射影几何和SAT求解器构建具有横向CCZ的码
Building codes with transversal CCZ using projective geometry and SAT solvers
arXiv 2610.09341 · 2026-10-08异构输入融合下的良性过拟合
Benign Overfitting under Heterogeneous Input Fusion
arXiv 2610.09340 · 2026-10-08序贯重置程序与错误发现率
Sequential resetting procedures and false discovery rate
arXiv 2610.09339 · 2026-10-08完全确定性的浅层神经网络与局部化激活函数。第一部分：构造与最优Sobolev逼近
Fully Deterministic Shallow Neural Networks with Localized Activations. Part I: Construction and Optimal Sobolev Approximation
arXiv 2610.09338 · 2026-10-08请教专家：基于LLM引导的强化学习用于自主网络防御
Ask the Expert: LLM-Guided Reinforcement Learning for Autonomous Cyber Defense
arXiv 2610.09337 · 2026-10-08多个黑洞的碰撞：数值相对论研究
Collisions of Multiple Black Holes: A Numerical Relativity Study
arXiv 2610.09336 · 2026-10-08SearchWorld：基于世界模型的无人机目标搜索空间价值想象
SearchWorld: Spatial Value-Grounded Imagination for UAV Object Search via World Models
arXiv 2610.09335 · 2026-10-08VM-ARRAYDPS：虚拟麦克风增强的扩散后验采样用于无监督盲语音分离
VM-ARRAYDPS: Virtual Microphone Augmented Diffusion Posterior Sampling for Unsupervised Blind Speech Separation
arXiv 2610.09334 · 2026-10-08统一辫群与Thompson群的Jones表示
On unifying Jones representations of braid groups and Thompson's group
arXiv 2610.09333 · 2026-10-08完全中心仿射极值超曲面与中心仿射Bernstein猜想
Complete centroaffine extremal hypersurfaces and centroaffine Bernstein conjecture
arXiv 2610.09332 · 2026-10-08投影提升定理、复反射群与Fermat簇的拟Galois点
Projective Lifting Theorem, Complex Reflection Groups, and Quasi-Galois Points for Fermat Varieties
arXiv 2610.09331 · 2026-10-08CRT-HMAR：面向开放任务感知的红外-可见光图像融合的因果需求追踪引导的分层多智能体调控
CRT-HMAR: Causal Requirement Tracing-Guided Hierarchical Multi-Agent Regulation for Open-Task-Aware Infrared-Visible Image Fusion
arXiv 2610.09330 · 2026-10-08GRC-Net：用于无监督多模态异常检测的全局表示一致性网络
GRC-Net: Global Representation Consistency Network for Unsupervised Multimodal Anomaly Detection
arXiv 2610.09329 · 2026-10-08Visual Jev 奖励：多主体图像生成的参考绑定验证
Visual Jev Rewards: Reference-Bound Verification for Multi-Subject Image Generation
arXiv 2610.09328 · 2026-10-08MCFR：一种用于鲁棒多型号板对板连接器组装的掩模引导的由粗到精回归框架
MCFR: A Mask-Guided Coarse-to-Fine Regression Framework for Robust Multi-Variant Board-to-Board Connector Assembly
arXiv 2610.09327 · 2026-10-08VIS-Ground：具有上下文锚定的视频交互式叙事
VIS-Ground: Video Interactive Storytelling with Contextual Grounding
arXiv 2610.09326 · 2026-10-08格点模拟中一阶相变的标量扰动与诱导引力波
Scalar Perturbations and Induced Gravitational Waves from First-Order Phase Transitions in Lattice Simulations
arXiv 2610.09325 · 2026-10-08带线性等式与不等式的随机投影
Random projection with linear equalities and inequalities
arXiv 2610.09324 · 2026-10-08数据稀缺星座任务中的约束扩散轨道蒙特卡洛方法
Constrained Diffusion for Data-Scarce Orbital Monte Carlo in Constellation Tasking
arXiv 2610.09323 · 2026-10-08开源量子网络模拟器的交叉验证
Cross-Validation of Open-Source Quantum Network Simulators
arXiv 2610.09322 · 2026-10-08方言鲁棒的语音语言模型与合成伪方言增强
Dialect-Robust Speech Language Models with Synthetic Pseudo-Dialect Augmentation
arXiv 2610.09321 · 2026-10-08理解对撞机事件流形的统一几何框架
A Unified Geometric Framework for Understanding Collider Event Manifolds
arXiv 2610.09320 · 2026-10-08带停止点的分次曲面的导出Picard重构
Derived Picard reconstruction for graded surfaces with stops
arXiv 2610.09319 · 2026-10-08CASM：用于节拍跟踪的上下文感知半马尔可夫后处理器
CASM: Context-Aware Semi-Markov Post-Processor for Beat Tracking
arXiv 2610.09318 · 2026-10-08SAREO-FM：用于SAR-EO基础模型的解耦语义监督
SAREO-FM: Decoupled Semantic Supervision for SAR-EO Foundation Models
arXiv 2610.09317 · 2026-10-08容许子图的结构性质与覆盖理想的符号幂的深度
Structural properties of admissible subgraphs and the depth of symbolic powers of cover ideals
arXiv 2610.09316 · 2026-10-08ClimbLab：腿式攀爬机器人MATLAB仿真平台
ClimbLab: MATLAB Simulation Platform for Legged Climbing Robotics
arXiv 2610.09315 · 2026-10-08$\omega$-稳定理论的类型$\omega$-Vaught猜想
The Type $ω$-Vaught's Conjecture for $ω$-stable Theories
arXiv 2610.09314 · 2026-10-08为什么视觉语言模型会漏检小目标，以及何时放大是安全的
Why VLMs Miss Small Objects, and When Zooming In Is Safe
arXiv 2610.09313 · 2026-10-08类型 $A_n$ 扇形坐标中簇单项式的对数凹性与单峰性
Log-concavity and unimodality of cluster monomials in fan coordinates of type $A_n$
arXiv 2610.09312 · 2026-10-08去噪块而非词元：基于分支词元实现的压缩连续扩散高效生成
Denoising Blocks, Not Tokens: Efficient Compressed Continuous Diffusion with Branching Token Realization
arXiv 2610.09311 · 2026-10-08各向异性膨胀的几何学用于最优正则化
The Geometry of Anisotropic Dilation for Optimal Regularization
arXiv 2610.09310 · 2026-10-08预测的未来还不够：学习机器人操作的可执行目标
Predicted Futures Are Not Enough: Learning Executable Goals for Robot Manipulation
arXiv 2610.09309 · 2026-10-08大型REBCO线圈中的分层
Delamination in Large REBCO Coils
arXiv 2610.09308 · 2026-10-08vLLM-Omni 技术报告：面向全模态生成的统一服务运行时
vLLM-Omni Technical Report: A Unified Serving Runtime for Omni-Modality Generation
arXiv 2610.09307 · 2026-10-08MovieSTAGE：用于电影fMRI ADHD分类的场景、转换与全局编码
MovieSTAGE: Scene, Transition, and Global Encoding for Movie-fMRI ADHD Classification
arXiv 2610.09306 · 2026-10-08Kuration SDK：通过数据整理弥合虚拟到现实差距
Kuration SDK: Addressing the Virtual2Real Gap via Data Curation
arXiv 2610.09305 · 2026-10-08emg2face：基于高密度表面肌电信号的表情动画生成
emg2face: Expressive Facial Animation with High-Density Surface EMG
arXiv 2610.09304 · 2026-10-08对数多个社区的半定规划问题的紧性与误差指数
Tightness and Error Exponents of SDP with Logarithmically Many Communities
arXiv 2610.09303 · 2026-10-08自适应风险认证的事件触发重规划用于动态导航
Adaptive Risk-Certified Event-Triggered Replanning for Dynamic Navigation
arXiv 2610.09302 · 2026-10-08一种具有补偿切向速度的能量稳定ALE-FEM方法用于Navier-Stokes自由边界流动
An energy--stable ALE--FEM with compensated tangential velocity for Navier--Stokes free--boundary flows
arXiv 2610.09301 · 2026-10-08关于其幂图均为良好覆盖的图
On graphs whose powers are well-covered
arXiv 2610.09300 · 2026-10-08无界域上 Hamilton--Jacobi--Bellman 方程的核方法与神经网络方法的收敛性
Convergence of kernel and neural-network methods for Hamilton--Jacobi--Bellman equations on unbounded domains
arXiv 2610.09299 · 2026-10-08节点级图神经架构搜索框架
Node-level Graph Neural Architecture Search Framework
arXiv 2610.09297 · 2026-10-08AI评估生态系统
The AI Evaluation Ecosystem
arXiv 2610.09296 · 2026-10-08
