2026-09论文中文摘要
J-Miner：从语言模型分类器中恢复可执行的决策知识
J-Miner: Recovering the Decision Logic of Fine-Tuned LLM Classifiers as Compact Rules
arXiv 2608.17063 · 2026-09-29民主的新数学
The New Mathematics of Democracy
arXiv 2608.16869 · 2026-09-29通过量子Littlewood-Richardson规则中的1-0松弛记录表与标记蜂巢中的格点，刻画n≤4时的𝔨最高权表与ĝ优表
Characterization of $\mathfrak{k}$-highest weight and $\widehat{\mathfrak{g}}$-dominant tableaux via $1$-$0$-slack recording tableaux in the quantum Littlewood-Richardson rule and lattice points in flagged hive polytopes
arXiv 2608.16800 · 2026-09-29梯度映射的离散性与开性的Pogorelov型反例
Critical Sobolev thresholds for openness and discreteness of Monge-Ampere gradient mappings
arXiv 2608.16757 · 2026-09-29扭曲的德利涅模与有限域上的阿贝尔簇
Twisted Deligne modules and abelian varieties over finite fields
arXiv 2608.16668 · 2026-09-29关于复多项式Bohnenblust--Hille常数的拟多项式上界
Polynomial growth of complex polynomial Bohnenblust--Hille constants
arXiv 2608.16584 · 2026-09-29操作约束下的覆盖最大化多项子集路由
Multinomial Subset Routing with Sum-Max Rewards and Operational Constraints
arXiv 2608.16375 · 2026-09-29AeroCopilotBench：用于在交互式虚拟驾驶舱环境中评估作为航空副驾驶的大语言模型智能体的双层基准
AeroCopilotBench: Safety-Gated Evaluation of LLM Agents on Aircraft Emergency Procedures in an Executable Cockpit
arXiv 2608.16349 · 2026-09-29SurgVIL：利用开源外科视频扩展外科机器人模仿学习
SurgVIL: Scaling Surgical Robot Imitation Learning with Open-source Surgical Videos
arXiv 2608.16058 · 2026-09-29Token分布与数据量：多领域会议摘要中的领域平衡
Token Distribution versus Data Volume: Domain Balancing in Multi-Domain Meeting Summarisation
arXiv 2608.15935 · 2026-09-29利用辅助系统辅助的投影纠缠对态（PEPS）获取带隙手性相
Bypassing the Chiral Obstruction in Two-Dimensional Tensor Networks
arXiv 2608.15732 · 2026-09-29SubZero+：基于大学习率的高效零阶大语言模型微调方法
SubZero+: Memory-Efficient Adaptive Zeroth-Order LLM Fine-Tuning in Random Subspaces
arXiv 2608.15665 · 2026-09-29用于机器遗忘的光谱显著性
Spectral Saliency for Machine Unlearning
arXiv 2608.15548 · 2026-09-29通过奇点范畴上的模研究有限维数
Finitistic dimension via modules over the singularity category
arXiv 2608.15482 · 2026-09-29SAGA：编码多表面用户动作序列的结构注意力生成动作嵌入模型
SAGA: Structure-Attended Generative Action Embedding Model that encodes Multi-Surface User Action Sequences
arXiv 2608.15429 · 2026-09-29分析说明：利用DELPHI开放数据的单电荷关联器
Analysis note: one-point charge correlator with DELPHI Open Data
arXiv 2608.15367 · 2026-09-29基于对决反馈的在线凸优化
Online Convex Optimization with Dueling Feedback
arXiv 2608.15050 · 2026-09-29预LN Transformer中部分残差消融的可复现性研究
A Reproducibility Study of Partial Residual Ablations in Pre-LN Transformers
arXiv 2608.14689 · 2026-09-29周期性环境胁迫塑造复杂生态网络的稳定性
Periodic Environmental Forcing Shapes the Stability of Complex Ecological Networks
arXiv 2608.14081 · 2026-09-29四维二次超曲面Q^4上的秩二弱法诺丛
Rank-two weak Fano bundles on a four-dimensional quadric hypersurface $Q^4$
arXiv 2608.13916 · 2026-09-29镁受限体积内的随机孪生：原位微力学测试与原子模拟的见解
Defect-controlled twin activation in crystallographically equivalent magnesium micropillars
arXiv 2608.13703 · 2026-09-29旋转QCD物质的临界行为与临界指数
Critical behavior and critical exponents of rotating QCD matter
arXiv 2608.13469 · 2026-09-29非周期性足以实现宏观热化
Aperiodicity is sufficient for macroscopic thermalization
arXiv 2608.13462 · 2026-09-29面向配备分布式储能的数据中心的概率功率供应理论
A Theory of Probabilistic Power Provisioning for Data Centers with Distributed Energy Storage
arXiv 2608.12993 · 2026-09-29DiCoR：用于高效遥感图像指称分割的解耦指称消歧与轮廓重校准方法
DiCoR: Decoupled Referent Disambiguation and Contour Recalibration for Efficient Referring Remote Sensing Image Segmentation
arXiv 2608.12980 · 2026-09-29AQuA：递归自我改进的量化交易研究智能体
AQuA: Recursively Self-Improving Quantitative Trading Research Agents
arXiv 2608.12841 · 2026-09-29ERSkill：面向技能引导的自适应记忆检索的演化框架
ERSkill: Evolving for Skill-Guided Adaptive Memory Retrieval
arXiv 2608.12720 · 2026-09-29面向大语言模型路由的误差感知反向拍卖机制
Error-Aware Reverse Auction Mechanism for Large Language Model Routing
arXiv 2608.12719 · 2026-09-29格点上有限对称性的费米子反常
Fermionic Anomalies of Finite Symmetries on Lattices
arXiv 2608.12455 · 2026-09-29从测试时缩放的视角理解在线策略蒸馏
Towards Understanding On-Policy Distillation through the Lens of Test-Time Scaling
arXiv 2608.11829 · 2026-09-29禁总大小尼姆游戏
Forbidden-Total-Size Nim
arXiv 2608.11665 · 2026-09-29FLARE++：带动态注意力路由的低秩注意力机制
FLARE++: Low-rank attention with dynamic attention routing
arXiv 2608.11519 · 2026-09-29基于广义利普希茨光滑性的神经网络梯度下降收敛性保证
Convergence Guarantees of Gradient Descent for Neural Networks via Generalized Lipschitz Smoothness
arXiv 2608.11479 · 2026-09-29基于显著性引导基元合并的紧凑前馈三维高斯模型
Compact Feed-Forward 3D Gaussians via Saliency-Guided Primitive Merging
arXiv 2608.10712 · 2026-09-29自动化与规模化针对AI智能体的行为科学研究
Causal Behavioral Evaluation of AI Agents at Scale via Automated Behavioral Science
arXiv 2608.10030 · 2026-09-29模型发现智能体：用于数据高效发现机制世界模型的大语言模型辅助贝叶斯实验设计
Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
arXiv 2608.09696 · 2026-09-29RecoverFly：面向空中视觉语言导航的故障感知强化学习后训练框架
RefineFly: Failure-Aware Post-Training for Aerial Vision-Language Navigation
arXiv 2608.09467 · 2026-09-29减少扩散语言模型中的预训练-生成不匹配
Reducing Pretraining-Generation Mismatch in Diffusion Language Models
arXiv 2608.09424 · 2026-09-29从目标到模型所学习的内容：不变学习的朗道理论
From Objectives to What Models Learn: A Landau Theory of Invariant Learning
arXiv 2608.09396 · 2026-09-29原生强化学习蒸馏：利用带分数轨迹实现少步图像生成
RL-Native Distillation: Exploiting Scored Trajectories for Few-Step Image Generation
arXiv 2608.09226 · 2026-09-29HOPPER：用于线性化图序列模型的可学习跳提取方法
HOPPER: Learnable Hop Extraction for Linearized Graph Sequence Models
arXiv 2608.09031 · 2026-09-29带物质的准拓扑引力的修正双副本构造
Modified Double Copy for Quasitopological Gravity with Matter
arXiv 2608.08988 · 2026-09-29搏动组织中神经线程放置的插入工具的基于预览的相对运动控制
Preview-Based Relative-Motion Control of an Insertion Tool for Neural-Thread Placement in Pulsating Tissue
arXiv 2608.08860 · 2026-09-29SDDBMs：软去噪扩散桥模型
SDDBMs: Soft Denoising Diffusion Bridge Models
arXiv 2608.08594 · 2026-09-29有限温度自旋适配的ROKS与TDDFT
Canonical-Ensemble Response from Spin-Adapted Time-Dependent Density Functional Theory
arXiv 2608.08454 · 2026-09-29Evidence-RL：面向证据密集型视觉推理
Evidence-RL: Towards Evidence-intensive Visual Reasoning
arXiv 2608.08021 · 2026-09-29跨批次边界的信息路由：Lipschitz 多臂老虎机中的内存-批次权衡
Memory--Batch Tradeoffs in Lipschitz Bandits
arXiv 2608.07922 · 2026-09-29利用Gevrey正则性打破量子偏微分方程求解器中的维度灾难
Breaking the Curse of Dimensionality in Quantum PDE Solvers via Gevrey Regularity
arXiv 2608.07893 · 2026-09-29Itô过程经验路径律的Sharp Wasserstein收敛速率
Sharp functional quantization and empirical Wasserstein rates for Itô processes
arXiv 2608.07879 · 2026-09-29视觉与WiFi融合：基于物理的体积力学特性估计
Vision Meets WiFi: Physics-Grounded Estimation of Volumetric Mechanical Properties
arXiv 2608.07726 · 2026-09-29
