2026-10论文中文摘要
CORE：面向可验证语言模型搜索的冲突导向推理消除
CORE: Conflict-Oriented Reasoning Elimination for Verifiable Language-Model Search
arXiv 2609.39069 · 2026-10-01SparseEngine：稀疏优先推理引擎
SparseEngine: Sparse-First Inference Engine
arXiv 2609.39068 · 2026-10-01Argus：一项关于预测Spot中断何时优于简单检查点的真实EKS研究
Argus: A Real-EKS Study of When Predicting Spot Interruptions Beats Simple Checkpointing
arXiv 2609.39067 · 2026-10-01智能体工具增强推理用于可解释的图像伪造检测
Agentic Tool-Augmented Reasoning for Explainable Image Forgery Detection
arXiv 2609.39066 · 2026-10-01代理能否信任其技能？揭示基于技能的LLM代理中的不安全信任链
Can Agents Trust Their Skills? Uncovering Unsafe Chains of Trust in Skill-Based LLM Agents
arXiv 2609.39065 · 2026-10-01冻结、验证、报告：基于公共证据审计城市站点规划
Freeze, Validate, Report: Auditing Urban Station Plans with Common Evidence
arXiv 2609.39064 · 2026-10-01毕达哥拉斯三元组2-正则性的一个初等证明
An elementary proof of the 2-regularity of Pythagorean triples
arXiv 2609.39063 · 2026-10-01突破有向哈密顿回路问题的 $2^n$ 障碍
Breaking the $2^n$ barrier for directed hamiltonicity
arXiv 2609.39062 · 2026-10-01跳跃回归再探
Jump regression revisited
arXiv 2609.39061 · 2026-10-01Autocata：基于能力中心型人工智能的反应网络引导的自主催化剂筛选
Autocata: Reaction-network-guided autonomous catalyst screening via capability-centric AI
arXiv 2609.39060 · 2026-10-01旋转曲面BRK型集合的尖锐体积界
Sharp volume bounds for BRK-type sets of surfaces of revolution
arXiv 2609.39059 · 2026-10-01面向RIS辅助超5G网络的3GPP合规基准数据集
A 3GPP-Compliant Benchmark Dataset for RIS-Aided Beyond 5G Networks
arXiv 2609.39058 · 2026-10-01双中子星系统的改进种群与核态方程联合推断
Improved Population and EOS Joint Inference for Binary Neutron Star Systems
arXiv 2609.39057 · 2026-10-01SteerQuant：在世界-动作模型中通过动作引导缩放来引导量化误差
SteerQuant: Steering Quantization Error with Action-Guided Scaling in World-Action Models
arXiv 2609.39056 · 2026-10-01缺失的系数：用于模型个性化的贝叶斯成对合并
The Missing Coefficients: Bayesian Pairwise Merging for Model Personalization
arXiv 2609.39055 · 2026-10-01CLIPPER 超越候选列表缩减：面向市政微出行政策变更的可审计决策支持
CLIPPER Beyond Shortlisting: Auditable Decision Support for Changing Municipal Micromobility Policies
arXiv 2609.39054 · 2026-10-01保结构傅里叶神经算子用于粗时间监督下的长时间Cahn-Hilliard动力学
Structure-preserving Fourier Neural Operators for Long-time Cahn-Hilliard Dynamics under Coarse Temporal Supervision
arXiv 2609.39053 · 2026-10-01一种具有改进竞争保证的在线运输鲁棒化贪心算法
A Robustified Greedy Algorithm for Online Transportation with Improved Competitive Guarantees
arXiv 2609.39052 · 2026-10-01TSMD：用于稳健视频精彩片段检测的时间-流模态丢弃
TSMD: Temporal-Stream Modality Dropout for Robust Video Highlight Detection
arXiv 2609.39051 · 2026-10-01隐蔽协助：多智能体系统中乐于助人的LLM智能体规避监管
Covert Assistance: Helpful LLM Agents Evade Oversight in Multi-Agent Systems
arXiv 2609.39050 · 2026-10-01结构 vs. 思维链：评估大语言模型在抑郁严重程度中的标准提取
Structure vs. Chain-of-Thought: Evaluating LLM Criteria Extraction for Depression Severity
arXiv 2609.39049 · 2026-10-01面向蛋白质定向进化的结构感知强化学习
Structure-aware Reinforcement Learning for Protein Directed Evolution
arXiv 2609.39048 · 2026-10-01BadAction：通过动作引导触发器对交互式视频生成进行后门攻击
BadAction: Backdoor Attacks on Interactive Video Generation via Action-Guided Triggers
arXiv 2609.39047 · 2026-10-01非线性奇异点传感中的定点分岔与路径信息界
Fixed-Point Bifurcations and Path-Information Bounds in Nonlinear Exceptional-Point Sensing
arXiv 2609.39046 · 2026-10-01RSIGame：具有递归自我改进的自主智能体游戏开发
RSIGame: Autonomous Agentic Game Development with Recursive Self-improvement
arXiv 2609.39045 · 2026-10-01游戏音效补全：基于事件级变换提示
Game Sound-Effect Completion with Event-Level Transformation Hints
arXiv 2609.39044 · 2026-10-01生成式与协同用户画像之间的路由：用于可控新颖性的服务时门控
Routing Between Generative and Collaborative User Profiles: A Serving-Time Gate for Controllable Novelty
arXiv 2609.39043 · 2026-10-01可证明的经典与量子局部算法用于Max-$k$-Cut及中等围长下的量子优势
Provable Classical and Quantum Local Algorithms for Max-$k$-Cut and Quantum Advantage at Moderate Girth
arXiv 2609.39042 · 2026-10-01神经节网络模型：多孔介质中截留相的演化
The Ganglion Network Model: Evolving Trapped Phases in Porous Media
arXiv 2609.39041 · 2026-10-01堆叠强透镜与弱透镜联合：改进测量红移$z\sim 0.5$处大质量早型星系的恒星与暗物质分布
Stacked strong and weak lensing united: Improved measurement of the stellar and dark matter distributions in massive early-type galaxies at $z\sim 0.5$
arXiv 2609.39040 · 2026-10-01幻影膜上的亚稳态暗能量
Metastable Dark Energy on the Phantom Brane
arXiv 2609.39039 · 2026-10-01回望以向前：生成式机器人策略的时间验证
Looking Back to Move Forward: Temporal Verification for Generative Robot Policies
arXiv 2609.39038 · 2026-10-01已部署验证器套件中的硬门候选资格
Hard-Gate Candidacy in a Deployed Validator Suite
arXiv 2609.39037 · 2026-10-01DESI 主巡天 DR1 中的“叉”形几何：目标选择对星系演化研究的影响
The "Fork" Geometry in the DESI Main Survey DR1: Implications of Target Selection for Galaxy Evolution Studies
arXiv 2609.39036 · 2026-10-01循环感知自编码器与跨信号一致性用于铁路车门异常检测
Cycle-Aware Autoencoder with Cross-SignalConsistency for Railway Door Anomaly Detection
arXiv 2609.39035 · 2026-10-01切换线性注意力
Switching Linear Attention
arXiv 2609.39034 · 2026-10-01TED：面向提示异常定位的文本轴证据分解
TED:Text-Axis Evidence Decomposition for Prompted Anomaly Localization
arXiv 2609.39033 · 2026-10-01预测MOS在复现人类对语音增强系统级偏好方面的可靠性如何？
How Reliable Are Predicted MOS for Reproducing Human System-Level Preferences in Speech Enhancement?
arXiv 2609.39032 · 2026-10-01$L^p$-算子代数的Morita等价性与étale群胚
Morita equivalence for $L^p$-operator algebras associated with étale groupoids
arXiv 2609.39031 · 2026-10-01SURE-EVAL：一个用于可复现评估的系统化统一智能体框架
SURE-EVAL: A Systematic and Unified Agentic Framework for Reproducible Evaluation
arXiv 2609.39030 · 2026-10-01胶子碎裂为$^1S_0^{[1,8]}$夸克偶素中阈值双对数的重求和
Resummation of threshold double logarithms in gluon fragmentation into $^1S_0^{[1,8]}$ quarkonia
arXiv 2609.39029 · 2026-10-01改进预测的MOS分数，而非感知质量：增强语音的多预测器测试时优化
Improving Predicted MOS Scores, Not Perceived Quality: Multi-Predictor Test-Time Optimization of Enhanced Speech
arXiv 2609.39028 · 2026-10-01可信 AI 审稿人缺失的一环：从修辞鲁棒性基准测试到 SciCore 审稿
A Missing Piece for Trustworthy AI Reviewers: From Benchmarking Rhetorical Robustness to SciCore Review
arXiv 2609.39027 · 2026-10-01搜索塑造结论：审计深度研究智能体中的证据选择偏差
Search Shapes Conclusions: Auditing Evidence Selection Bias in Deep Research Agents
arXiv 2609.39026 · 2026-10-01算法公平性的秩毕业度量
A Rank Graduation metric for Algorithmic fairness
arXiv 2609.39025 · 2026-10-01文本到图像模型的持久水印
Persistent Watermarking of Text-to-Image Models
arXiv 2609.39024 · 2026-10-01当注意力护栏成为学习的障碍：迈向临界点
When Attention Guardrails Become Barriers to Learning: Towards the Tipping Point
arXiv 2609.39023 · 2026-10-01视频推理的帧差分在线策略自蒸馏
Frame Differential On-Policy Self-Distillation for Video Reasoning
arXiv 2609.39021 · 2026-10-01深度ReLU神经网络学习谱Barron函数的极小极大速率
Minimax rates for learning spectral Barron functions by deep ReLU neural networks
arXiv 2609.39020 · 2026-10-01同步多视角神经扩散
Synchronous Multi-view Neural Diffusion
arXiv 2609.39019 · 2026-10-01
