2026-10论文中文摘要
不要重复自己：面向覆盖率的自监督微调
Don't Repeat Yourself: Self-Supervised Fine-Tuning for Coverage
arXiv 2609.31688 · 2026-10-01语言增强的视频动作预测：设计基础、基准与开放挑战
Language-Augmented Video Action Anticipation: Design Fundamentals, Benchmarks, and Open Challenges
arXiv 2609.31665 · 2026-10-01面向联邦数据的可验证隐私学习
Toward provably private learning from federated data
arXiv 2609.31494 · 2026-10-01带状态约束的最优控制中的奇异倒向随机微分方程
Singular Backward SDEs for Optimal Control with State Constraints
arXiv 2609.31487 · 2026-10-01AxonSynth：用于光片显微镜中零样本3D轴突分割的域随机化合成数据
AxonSynth: Domain-Randomized Synthetic Data for Zero-Shot 3D Axon Segmentation in Light-Sheet Microscopy
arXiv 2609.31431 · 2026-10-01广义中心三项式系数的参数化 Catalan 型同余
A parametric Catalan-type congruence for generalized central trinomial coefficients
arXiv 2609.31423 · 2026-10-01分位数曲面的渐近形状
On the asymptotic shape of quantile surfaces
arXiv 2609.31345 · 2026-10-01液晶中短程序参量凝聚导致的纳米尺度调制的涌现
Emergence of Nanoscale Modulation in Liquid Crystals as a Result of Short-Range Order Parameter Condensation
arXiv 2609.30916 · 2026-10-01LLM帕金森综合征：执行控制失败、令牌低效持续行为，以及面向自主语言模型代理的不确定性感知全局执行控制架构
LLM Parkinsonism: Executive-Control Failure, Token-Inefficient Persistence, and an Uncertainty-Aware Global Executive Control Architecture for Autonomous Language-Model Agents
arXiv 2609.30662 · 2026-10-01HuGo：大型语言模型作为人形机器人全身策略代码设计器用于移动操作
HuGo: LLMs as Whole-Body Policy Code Designers for Humanoid Loco-Manipulation
arXiv 2609.30594 · 2026-10-01耗散孤子共振区弱光注入对噪声辅助多脉冲的能量选择性控制
Energy-selective control of noise-assisted multipulsing by weak optical seeding in the dissipative-soliton-resonance regime
arXiv 2609.30522 · 2026-10-01RAZOR：剪除大语言模型中可替换的专家
RAZOR: Pruning Replaceable Experts in LLMs
arXiv 2609.30465 · 2026-10-01信息论与代数重整化群
Informational and algebraic renormalization group
arXiv 2609.30362 · 2026-10-01非阿贝尔量子元胞自动机：$1{+}1$ 维 $SU(2)$ 杨-米尔斯理论（含费米子）
Non-abelian quantum cellular automata: $1{+}1$-dimensional $SU(2)$ Yang--Mills with fermions
arXiv 2609.30285 · 2026-10-01形成纠缠的Moreau-Yosida近似：基本性质与精度估计
The Moreau-Yosida approximation of the Entanglement of Formation: basic properties and accuracy estimates
arXiv 2609.30246 · 2026-10-01气候、科学与FCC：清晰思考、明智选择、迅速行动
Climate, Science, and the FCC: Think Clearly, Choose Wisely, Act Swiftly
arXiv 2609.30180 · 2026-10-01MQSS-Selector：用于MLIR编译管线的强化学习引导的Pass选择
MQSS-Selector: RL-Guided Pass Selection for an MLIR Compilation Pipeline
arXiv 2609.30104 · 2026-10-01由单项式理想商群的线性形式构成的参数系统
Systems of parameters consisting of linear forms for monomial ideal quotients
arXiv 2609.30081 · 2026-10-01目标放近，方能致远：冻结的世界模型比你想的更能规划
Aim Short to Reach Far: Your Frozen World Model Can Plan Better Than You Think
arXiv 2609.30036 · 2026-10-01对抗性影响在多智能体系统中如何扩展？
How does Adversarial Influence Scale in Multi-Agent Systems?
arXiv 2609.30028 · 2026-10-01宇宙学常数的红外屏蔽
Infrared Screening of the Cosmological Constant
arXiv 2609.29979 · 2026-10-01利用Fisher信息对潜在场进行快速不确定性量化
Rapid Uncertainty Quantification on a Latent Field using Fisher Information
arXiv 2609.29969 · 2026-10-01强制性条件与广义Monge-Ampère方程
The coerciveness and generalised Monge-Ampère equations
arXiv 2609.29968 · 2026-10-01硬停止：针对恶意自主执行的内核级抢占与遏制
Hard Stop: Kernel-Level Preemption and Containment for Rogue Agentic Execution
arXiv 2609.29808 · 2026-10-01PPTBench：编码智能体能否通过结构化、可编辑的幻灯片重建视觉世界？
PPTBench: Can Coding Agents Reconstruct the Visual World through Structured, Editable Slides
arXiv 2609.29718 · 2026-10-01经典测量理论误导LLM评判者的三种方式
Three Ways Classical Test Theory Can Mislead About LLM Judges
arXiv 2609.29709 · 2026-10-01持续学习的序列代价
The Sequential Price of Continual Learning
arXiv 2609.29674 · 2026-10-01KAYROS：一种用于带时间窗的时变车辆路径问题的任意时间精确求解器 - 扩展版
KAYROS: An Anytime and Exact Solver for the Time-Dependent Vehicle Routing Problem with Time Windows - Extended Version
arXiv 2609.29597 · 2026-10-01ATLAS：用于自动化结构的原子翻译与语言
Building Atomic Structures from Natural Language
arXiv 2609.29595 · 2026-10-01爱因斯坦引力与内部相互作用的统一
Unification of Einstein Gravity with Internal Interactions
arXiv 2609.29590 · 2026-10-01自相似集与测度的绝对连续性和维数守恒
Absolute continuity of projections and dimension conservation for self-similar sets and measures
arXiv 2609.29301 · 2026-10-01黑洞度规与引力阴影的重建
Reconstruction of Black Hole Metric with Gravitational Shadows
arXiv 2609.29036 · 2026-10-01CrossSafe：迈向跨具身潜在安全滤波器
CrossSafe: Towards Cross-Embodiment Latent Safety Filters
arXiv 2609.28984 · 2026-10-01可重构基于总线的量子路由器用于模块化超导处理器
Reconfigurable Bus-based Quantum Router for Modular Superconducting Processors
arXiv 2609.28881 · 2026-10-01基于任务的拉曼光谱质量度量评估框架
Validating Spectral Quality Measures as Proxies for Task Performance: A Controlled-Perturbation Framework for Raman Spectroscopy
arXiv 2609.28874 · 2026-10-01人在回路地理空间标注：面向现场部署无人机系统的快速数据集构建
Human-in-the-Loop Geospatial Annotation for Rapid Dataset Construction in Field-Deployed UAV Systems
arXiv 2609.28767 · 2026-10-01r-安全余量与自适应运动平滑用于跟车过渡
The r-Safety Reserve and Adaptive Kinematic Smoothing for Car-Following Transitions
arXiv 2609.28676 · 2026-10-01当评估者喊“狼来了”：生产环境中教育AI的LLM即评估者评估经验教训
When Evaluators Cry Wolf: Lessons from Production LLM-as-Judge Evaluation in Educational AI
arXiv 2609.28478 · 2026-10-01从预测到可解释的提供者行为画像，用于欺诈、浪费和滥用审查
From Prediction to Explainable Provider Behavior Profiles for Fraud, Waste, and Abuse Review
arXiv 2609.28477 · 2026-10-01StudentBench：AI与人类辅导在GRE学习收益上效果相当
StudentBench: AI and human tutoring yield equivalent GRE learning gains
arXiv 2609.28470 · 2026-10-01数学推理中的顺序不变答案与顺序敏感表示
Order-Invariant Answers, Order-Sensitive Representations in Mathematical Reasoning
arXiv 2609.28442 · 2026-10-01利用可微高斯表示学习集体动力学
Learning Collective Dynamics with Differentiable Gaussian Representations
arXiv 2609.28405 · 2026-10-01OranSim：模拟社交媒体营销
OranSim: Simulating Consumer Response to Social Media Campaigns Before Launch
arXiv 2609.28388 · 2026-10-01RL始于RL之前：关于策略蒸馏以改进强化学习
RL Starts before RL: On Policy Distillation for Better Reinforcement Learning
arXiv 2609.28145 · 2026-10-01混合特征下通过叠的局部解析表示
Locally analytic representations in mixed characteristic via stacks
arXiv 2609.28141 · 2026-10-01周期哈密顿近似与\mathbb{C^2}中的Oka补集
Periodic transport and Oka complements in \(\C^n\)
arXiv 2609.28089 · 2026-10-01本质上非有限基的加法幂等半环
Inherently nonfinitely based additively idempotent semirings
arXiv 2609.27641 · 2026-10-01大型知识模型：从论文到科学推理景观
Large Knowledge Model: A Knowledge Foundation for Agentic Science at Scale
arXiv 2609.27297 · 2026-10-01SR-Fraud：一种面向非平稳支付欺诈检测的结果监督反思式LLM智能体框架
SR-Fraud: An Outcome-Supervised Reflective LLM Agent Framework for Non-Stationary Payment Fraud Detection
arXiv 2609.27287 · 2026-10-01NaviScale：生成用于目标导航的大规模语义地图数据集
NaviScale: Generating Large-Scale Semantic Map Datasets for Object Navigation
arXiv 2609.27218 · 2026-10-01
