2026-09论文中文摘要
探索基于生成式智能体模型的因果机制
Exploring Causal Mechanisms with Generative Agent-Based Models
arXiv 2609.35819 · 2026-09-30IMPACT：面向云边系统中SLO保障的微服务迁移的意图驱动多智能体策略
IMPACT: Intent-driven Multi-agent Policy with Attention for SLO-guaranteed Microservice Migration in Cloud-edge Systems
arXiv 2609.35818 · 2026-09-30更少均匀的离散扩散更强大且可扩展
Less Uniform Discrete Diffusion is More Powerful and Scalable
arXiv 2609.35817 · 2026-09-30PrimeSeeker：面向深度搜索智能体的能力导向监督
PrimeSeeker: Capability-Oriented Supervision for Deep Search Agents
arXiv 2609.35816 · 2026-09-30如何在LLM裁判上运行统计并信任结果：使用evalstats对小样本AI评估进行校准推断
How to Run Statistics over LLM Judges and Trust the Results: Calibrated Inference for Small-Sample AI Evaluation with evalstats
arXiv 2609.35815 · 2026-09-30通过受控环境干预构建具有挑战性的浏览器使用任务
Constructing Challenging Browser-Use Tasks by Controlled Environment Interventions
arXiv 2609.35814 · 2026-09-30LLM智能体社会中的局部可预测性与集体保真度
Local Predictability and Collective Fidelity in LLM-Agent Societies
arXiv 2609.35813 · 2026-09-30车载对话助手中多轮对话的自动化评估
Automated Evaluation of Multi-Turn Dialogues in In-Car Conversational Assistants
arXiv 2609.35812 · 2026-09-30Lookahead-R：通过执行中心规划进行预算感知的工具检索
Lookahead-R: Budget-Aware Tool Retrieval via Execution-Centric Planning
arXiv 2609.35811 · 2026-09-30TRACE：面向肿瘤学大语言模型的可部署树关系结构增强框架
TRACE: Deployable Tree-Relational Structure Enhancement for Oncology LLMs
arXiv 2609.35810 · 2026-09-30多模态大语言模型能否生成和检测多模态社交媒体假新闻？
Can Multimodal Large Language Models Generate and Detect Multimodal Social Media Fake News?
arXiv 2609.35809 · 2026-09-30当成功记忆误导具身智能体：面向任务条件执行的内存自适应
When Successful Memories Mislead Embodied Agents:Memory Adaption For Task-Conditioned Execution
arXiv 2609.35808 · 2026-09-30环境引导：利用数据流控制提升智能体效用与安全性
Environment Steering: Using Data Flow Control to Improve Agent Utility and Safety
arXiv 2609.35807 · 2026-09-30从词汇基线到智能体检索增强生成：基于SFIA框架的结构化技能与责任级别提取
From Lexical Baselines to Agentic Retrieval-Augmented Generation: Structured Skill and Responsibility-Level Extraction with the SFIA Framework
arXiv 2609.35806 · 2026-09-30评估提示扰动对大型语言模型中偏见与幻觉的影响
Evaluating the Effects of Prompt Perturbation on Bias and Hallucination in Large Language Models
arXiv 2609.35804 · 2026-09-30近似具有任意成本的组合合约
Approximating Combinatorial Contracts with Arbitrary Costs
arXiv 2609.35803 · 2026-09-30基数约束下的市场份额均衡随机商品组合
Cardinality-Constrained Randomized Assortments with Balanced Market Share
arXiv 2609.35802 · 2026-09-30精确Hill份额是同时保证
Exact Hill Shares Are Simultaneous Guarantees
arXiv 2609.35801 · 2026-09-30HeadGuard：面向低比特VLM KV缓存量化的选择性头部保护
HeadGuard: Selective Head Protection for Low-Bit VLM KV-Cache Quantization
arXiv 2609.35800 · 2026-09-30OpenAI-HuggingFace：复现与对齐测试的经验教训
OpenAI-HuggingFace: A Reproduction & Lessons for Alignment Testing
arXiv 2609.35799 · 2026-09-30相似度图统计的最优遍数与完美采样
Optimal Passes and Perfect Sampling for Similarity Graph Statistics
arXiv 2609.35798 · 2026-09-30二值化展平分数空间
Binarization Flattens the Score Space
arXiv 2609.35797 · 2026-09-30开发一个用于提取韩语发票信息的OCR模型
Developing an OCR model for Extracting Information from Invoices with Korean Language
arXiv 2609.35796 · 2026-09-30校准优先的跨队列多模态时序学习用于可迁移的哮喘风险预测
Calibration-First Cross-Cohort Multimodal Temporal Learning for Transferable Asthma-Risk Forecasting
arXiv 2609.35795 · 2026-09-30筛与智：面向可靠 RALM 弃权的高效干扰过滤
Sieve and Sage: Efficient Distraction Filtering for Reliable RALM Abstention
arXiv 2609.35794 · 2026-09-30LSTM故障检测器的无服务器八卦训练：与联邦、本地及集中式学习在NASA C-MAPSS上的匹配协议比较
Serverless gossip training of LSTM failure detectors: A matched-protocol comparison with federated, local and centralized learning on NASA C-MAPSS
arXiv 2609.35792 · 2026-09-30FD-VAD：面向流式全双工语音的语义端点检测
FD-VAD: Semantic Endpoint Detection for Streaming Full-Duplex Speech
arXiv 2609.35791 · 2026-09-30Agent可调用功能覆盖率：衡量软件对AI Agent的就绪度
Agent-Callable Feature Coverage: Measuring Software Readiness for AI Agents
arXiv 2609.35789 · 2026-09-30用于等几何分析的局部修改非钳制片段的可容许样条空间的混合重构
Hybrid Reconstruction of Admissible Spline Spaces from Locally Modified Unclamped Patches for Isogeometric Analysis
arXiv 2609.35788 · 2026-09-30预印本演变：arXiv 上未审稿草稿的兴起及其对天文学的影响
The Preprint Evolution: The Rise of Unreviewed Drafts on arXiv and its Implications for Astronomy
arXiv 2609.35787 · 2026-09-30多智能体LLM对话的定向可追溯调查：基于知识图谱的语义捆绑
Targeted and Traceable Investigation of Multi-Agent LLM Dialogue via Semantic Bundling of Knowledge Graphs
arXiv 2609.35786 · 2026-09-30中子探测器慢化效率：蒙特卡罗模拟、实验验证与角度依赖性
The efficiency of moderating neutron detector: Monte Carlo simulation, experimental validation, and angular dependence
arXiv 2609.35785 · 2026-09-30软课程学习用于优化新鲜与泛化推荐
Soft Curriculum Learning for Optimizing Fresh and Generalized Recommendations
arXiv 2609.35783 · 2026-09-30金融证据拥挤：检索增强生成中约束诱导位移的诊断与缓解
Financial Evidence Crowding: Diagnosing and Mitigating Constraint-Induced Displacement in Retrieval-Augmented Generation
arXiv 2609.35782 · 2026-09-30发现（与遗漏）算法偏见：调查公平性评估工具中的用户理解
Spotting (and Missing) Algorithmic Bias: Investigating User Understanding in a Fairness Assessment Tool
arXiv 2609.35781 · 2026-09-30TSG Suggester：面向云事件管理的故障排除指南推荐中的树状结构知识图谱检索
TSG Suggester: Tree-Structured Knowledge-Graph Retrieval for Troubleshooting Guide Recommendation in Cloud Incident Management
arXiv 2609.35780 · 2026-09-30大型语言模型表现出类似人类的贝叶斯虚伪
Large Language Models Exhibit Human-Like Bayesian Hypocrisy
arXiv 2609.35779 · 2026-09-30基于单次试验脑电的在线人类意图推断作为潜在控制状态
Online Inference of Human Intention as a Latent Control State from Single-Trial EEG
arXiv 2609.35778 · 2026-09-30SensWear：一个开放、模块化且AI就绪的可穿戴平台
SensWear: An Open, Modular, and AI-Ready Wearable Platform
arXiv 2609.35777 · 2026-09-30教师视角下基于AI的多智能体仿真设计以应对校园欺凌
Teachers' perspective on AI-based Multi-Agent Simulation Design to Combat School Bullying
arXiv 2609.35776 · 2026-09-30SPECTRA：面向自主边缘-云GUI接地（GUI Grounding）的端侧认知扰动与轨迹分析
SPECTRA: On-Device Cognitive Perturbation and Trajectory Analysis for Autonomous Edge-Cloud GUI Grounding
arXiv 2609.35775 · 2026-09-30生成后验证主导检索优化：RAG 流水线特征的 2^4 全因子消融实验
Post-Generation Verification Dominates Retrieval Optimization: A 2^4 Factorial Ablation of RAG Pipeline Features
arXiv 2609.35774 · 2026-09-30Socrates-RAG：针对协同证据投毒的基于前提的主动检索
Socrates-RAG: Premise-Directed Inquiry against Coordinated Evidence Poisoning
arXiv 2609.35773 · 2026-09-30三个立方满数之和的密度亏缺
A density deficit for sums of three cube-full numbers
arXiv 2609.35772 · 2026-09-30TokenCast：预测LLM智能体执行过程中的令牌消耗
TokenCast: Forecasting Token Consumption During LLM Agent Execution
arXiv 2609.35760 · 2026-09-30KV-streams：用于智能体强化学习中高效压缩的键值流
KV-streams for Efficient Compaction in Agentic Reinforcement Learning
arXiv 2609.35750 · 2026-09-30GeoVerse：几何潜在空间中的世界一致新视角合成
GeoVerse: World-Consistent Novel View Synthesis in Geometric Latent Space
arXiv 2609.35734 · 2026-09-30快速射电暴-持续射电源系统III：PRS光度与FRB旋转测量之间的关系
Fast radio burst - persistent radio source systems III. The relation between PRS luminosity and FRB rotation measure
arXiv 2609.35733 · 2026-09-30RoboCompiler：闭链机器人的图原生编译，实现一致的建模、控制与仿真
RoboCompiler: Graph-Native Compilation of Closed-Chain Robots for Consistent Modeling, Control, and Simulation
arXiv 2609.35717 · 2026-09-30基于连续潜在扩散的推理
Reasoning with Continuous Latent Diffusion
arXiv 2609.35694 · 2026-09-30
