2026-08论文中文摘要
当策略改变概率时：用于LLM代码审查的模块化决策
When Policies Change Probabilities: Modular Decision-Making for LLM Code Review
arXiv 2608.02677 · 2026-08-05开普勒轨道与牛顿平方反比定律等价性的构造性欧几里得证明
Constructive Euclidean Proofs of the Equivalence Between Keplerian Orbits and Newton's Inverse-Square Law
arXiv 2608.02676 · 2026-08-05关于Erdős-Gyárfás猜想的立方二分反例的60顶点下界
A 60-Vertex Lower Bound for Cubic Bipartite Counterexamples to the Erdős-Gyárfás Conjecture
arXiv 2608.02675 · 2026-08-05移动安全屏障：针对白盒攻击的动态路由自适应对齐
Moving the Safety Barrier: Dynamic Routing Adaptive Alignment Against White-Box Attacks
arXiv 2608.02674 · 2026-08-05安全优先的Text-to-Terraform评估：针对安全基础设施即代码生成的大语言模型（LLMs）与小型语言模型（SLMs）基准测试
Security-First Evaluation of Text-to-Terraform: Benchmarking LLMs and SLMs for Secure IaC Generation
arXiv 2608.02672 · 2026-08-05基于硬件性能计数器的恶意软件检测分类器性能研究
On the Performance of Malware Detection Classifiers Using Hardware Performance Counters
arXiv 2608.02671 · 2026-08-05权限受限：强化环境下编码智能体的策略分级评估
Permission Denied: Policy-Graded Evaluation of Coding Agents in Hardened Environments
arXiv 2608.02670 · 2026-08-05曝光度最高的 Docker Hub 镜像中的漏洞、秘密与配置错误
Vulnerabilities, Secrets and Misconfiguration in the Highest-Exposure Docker Hub Images
arXiv 2608.02669 · 2026-08-05球面收缩归一化
Sphere Retraction Normalizations
arXiv 2608.02668 · 2026-08-05重建初创企业：远程工作与技能互补性的实证研究
Rebuilding Startups: An Empirical Study on Remote Work and Skill Complementarity
arXiv 2608.02666 · 2026-08-05ZK-SR117：面向聚合公平借贷指标的分块零知识证明设计，及指向完整SR 11-7覆盖的控制映射
ZK-SR117: A Chunked Zero-Knowledge Attestation Design for Aggregated Fair-Lending Metrics, with a Control Mapping toward Full SR 11-7 Coverage
arXiv 2608.02664 · 2026-08-05CT-HEG：用于ICU院内死亡率预测的带时间戳属性的双向事件图——一项架构消融研究
CT-HEG: A Bidirectional, Timestamp-Attributed Event Graph for ICU In-Hospital Mortality Prediction - An Architectural Ablation Study
arXiv 2608.02663 · 2026-08-05基于预训练符号 Transformer 的物理动力系统验证器引导式模型发现
Verifier-Guided Model Discovery for Physical Dynamical Systems with Pretrained Symbolic Transformers
arXiv 2608.02662 · 2026-08-05超越不变字典：基于滤波扩展动态模态分解的数据驱动Koopman谱恢复
Beyond Invariant Dictionary: Data-Driven Koopman Spectral Recovery with Filtered Extended Dynamic Mode Decomposition
arXiv 2608.02661 · 2026-08-05动态限应变弹性力学中拟线性双曲方程的时空有限元近似
Space-Time Finite Element Approximation of Quasilinear Hyperbolic Equations Arising in Dynamic Strain-Limiting Elasticity
arXiv 2608.02658 · 2026-08-05面向多租户云平台知识产权保护的安全AI水印框架
Secure AI Watermarking Framework for IP Protection in Multi-Tenant Cloud Platforms
arXiv 2608.02656 · 2026-08-05规范不变洛伦兹破缺质量矢量场的重叠共振分支
Overlapping resonance branches of a gauge-invariant Lorentz-violating massive vector
arXiv 2608.02655 · 2026-08-05面向自动证明论语义：三维K3与LP的推理-行为语义
Towards Automated Proof-Theoretic Semantics: Inference-Behaviour Semantics for 3-Dimensional K3 and LP
arXiv 2608.02654 · 2026-08-05Light-Loco-Parkour：通过多技能蒸馏实现通用的感知型全身运动
Light-Loco-Parkour: Versatile Perceptive Whole-Body Locomotion via Multi-Skill Distillation
arXiv 2608.02653 · 2026-08-05公私部门信贷项目中的委托监督：投资不足、投资过度及补贴贷款的设计
Delegated Monitoring in Public-Private Sector Credit Programs: Underinvestment, Overinvestment, and the Design of Subsidized Lending
arXiv 2608.02651 · 2026-08-05HyperAgent：面向工具-模式超图的规划与执行，用于工具使用型大语言模型智能体
HyperAgent: Planning and Acting over Tool-Schema Hypergraphs for Tool-Use LLM Agents
arXiv 2608.02650 · 2026-08-05带断言引导执行与可检查中间表示的意图级量子编程
Intent-Level Quantum Programming with Assertion-Guided Execution and Inspectable Intermediate Representation
arXiv 2608.02648 · 2026-08-05用于三份Werner蒸馏的Sharp Plucker几何
Sharp Plucker Geometry for Three-Copy Werner Distillation
arXiv 2608.02647 · 2026-08-05跨麻醉剂ECoG状态解码在决策阈值处失败，而非表征层面
Cross-Anesthetic ECoG State Decoding Fails at the Decision Threshold, Not the Representation
arXiv 2608.02646 · 2026-08-05经验证的工具调用可提升非原子故障下LLM智能体的可靠性
Verified Tool Calls Improve LLM Agent Reliability Under Non-Atomic Failures
arXiv 2608.02645 · 2026-08-05用于微服务识别与重构的统一特征模型
A Unified Feature Model for Microservice Identification and Refactoring
arXiv 2608.02644 · 2026-08-05MDArena：面向真实分子动力学工作流的编码智能体评估基准
MDArena: Evaluating Coding Agents on Realistic Molecular Dynamics Workflows
arXiv 2608.02642 · 2026-08-05IR2Solve：面向成本高效优化自动建模的结构化中间表示
IR2Solve: Structured Intermediate Representations for Cost-Efficient Optimization Autoformulation
arXiv 2608.02641 · 2026-08-05指令堆叠崩溃：基准测试集与提示词编译的能力依赖价值
Instruction Stacking Collapse: A Benchmark and the Capability-Dependent Value of Prompt Compilation
arXiv 2608.02639 · 2026-08-05研究、识别与修复人工智能密集型网络物理系统中的隐性技术债务
Studying, Identifying, and Fixing Hidden Technical Debt in AI-Intensive Cyber-Physical Systems
arXiv 2608.02638 · 2026-08-05分子构象光谱的有限分辨率可识别性与测量设计
Finite-Resolution Identifiability and Measurement Design for Molecular Conformer Spectroscopy
arXiv 2608.02637 · 2026-08-05反思自进化智能体技能：多轮次的反馈动态
Rethinking Self-Evolving Agent Skills: Feedback Dynamics over Multiple Rounds
arXiv 2608.02636 · 2026-08-05维格纳的朋友悖论再探讨
Wigner's Friend Paradox Revisited
arXiv 2608.02635 · 2026-08-05特征2下可分雅可比猜想的二维反例
A Dimension-Two Counterexample to the Separable Jacobian Conjecture in Characteristic Two
arXiv 2608.02634 · 2026-08-05GeoID-PINN：考虑地理耦合的可识别性感知区域流行病推断方法
GeoID-PINN: Identifiability-Aware Regional Epidemic Inference with Geographic Coupling
arXiv 2608.02633 · 2026-08-05大型语言模型（LLMs）可对归因图进行标注
LLMs Can Annotate Attribution Graphs
arXiv 2608.02632 · 2026-08-05图中的超k-安全集
On Ultra $k$-Secure Sets in Graphs
arXiv 2608.02631 · 2026-08-05PULSE：用于时空知识图谱工程的可执行合约语言
PULSE: An Executable Contract Language for Spatiotemporal Knowledge Graph Engineering
arXiv 2608.02630 · 2026-08-05用于建模地质碳封存中可变操作与量化不确定性的多模态自回归Transformer代理模型
Multimodal Auto-regressive Transformer Surrogate for Modeling Variable Operations and Quantifying Uncertainty in Geological Carbon Storage
arXiv 2608.02629 · 2026-08-05零信任软件定义网络架构中的微分段异常检测
Micro-Segmentation Anomaly Detection in Zero-Trust Software-Defined Network Fabrics
arXiv 2608.02627 · 2026-08-05分层极小集附近的拉普拉斯渐近行为
Laplace Asymptotics near Stratified Minimum Sets
arXiv 2608.02626 · 2026-08-05推测校正：扩散语言模型的草稿后精炼解码
Speculative Correction: Draft-then-Refine Decoding for Diffusion Language Models
arXiv 2608.02625 · 2026-08-05基于双图册选择器的二元细化迭代的神经网络实现
Neural network realization of binary refinement iterates via a two-chart atlas selector
arXiv 2608.02624 · 2026-08-05解决气泡谜题：微气泡中过氧化氢的形成先于羟基自由基，且受固-水界面调控
Resolving the Bubble Puzzle: Hydrogen Peroxide Formation Precedes Hydroxyl Radicals in Microbubbles and is Governed by Solid-Water Interfaces
arXiv 2608.02622 · 2026-08-05知晓形式，而非功能：自动审计法律基准中答案与权威的脱钩问题
Knowing the Form, Not the Function: Automatically Auditing Answer--Authority Decoupling in Legal Benchmarks
arXiv 2608.02621 · 2026-08-05JudgeArena：用于可复现LLM-评判者评估的统一框架
JudgeArena: A Unified Framework for Reproducible LLM-Judge Evaluation
arXiv 2608.02620 · 2026-08-05刺激系统的公理式共享介质协调
Axiomatic shared-medium coordination for stigmergic systems
arXiv 2608.02619 · 2026-08-05超越群体思维：通过元角色锚定与顺序温度缩放规避大语言模型同质化
Beyond the Hivemind: Escaping LLM Homogeneity via Meta-Persona Anchoring and Sequential Temperature Scaling
arXiv 2608.02618 · 2026-08-05OncoTriad-QA：用于泛癌推理的患者级放射学-病理学-基因组学基准
OncoTriad-QA: A Patient-Level Radiology-Pathology-Genomics Benchmark for Pan-Cancer Reasoning
arXiv 2608.02615 · 2026-08-05结构稀疏图中最大团枚举的近最优算法
Near-Optimal Algorithms for Maximal Clique Enumeration in Structurally Sparse Graphs
arXiv 2608.02614 · 2026-08-05
