cs.AI（2026-05-04）

📊 共 31 篇论文 | 🔗 4 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (26 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (5 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (26 篇)

#题目一句话要点标签🔗⭐
1 Foundation-Model-Based Agents in Industrial Automation: Purposes, Capabilities, and Open Challenges 综述研究：基于大模型的智能体在工业自动化中的应用、能力与挑战 large language model foundation model
2 Position: How can Graphs Help Large Language Models? 图结构助力大语言模型：提升知识、推理与结构化数据理解能力 large language model chain-of-thought
3 On Training Large Language Models for Long-Horizon Tasks: An Empirical Study of Horizon Length 研究长程任务中LLM训练，揭示任务长度对训练稳定性和泛化性的影响 large language model
4 When Audio-Language Models Fail to Leverage Multimodal Context for Dysarthric Speech Recognition 针对构音障碍语音识别，研究表明现有语音-语言模型未能有效利用多模态临床上下文信息。 multimodal
5 Foundation Models to Unlock Real-World Evidence from Nationwide Medical Claims 提出ReClaim：基于大规模医疗理赔数据的医疗健康领域预训练模型 foundation model
6 ProPACT: A Proactive AI-Driven Adaptive Collaborative Tutor for Pair Programming ProPACT：用于结对编程的主动式AI驱动自适应协作辅导系统 multimodal
7 Anon: Extrapolating Optimizer Adaptivity Across the Real Spectrum 提出Anon优化器，通过可调适应性和增量延迟更新，统一并超越经典与现代优化器。 large language model
8 Submodular Benchmark Selection 提出基于次模优化的基准测试选择方法，降低大模型评测成本。 large language model
9 AI-Generated Smells: An Analysis of Code and Architecture in LLM and Agent-Driven Development 揭示AI生成软件的技术债务：推理复杂度与代码质量的权衡分析 large language model
10 Hybrid Inspection and Task-Based Access Control in Zero-Trust Agentic AI 提出混合检查与任务型访问控制，保障零信任Agentic AI安全。 large language model
11 Trustworthy AI Suffers from Invariance Conflicts and Causality is The Solution 利用因果关系解决可信AI中不变性冲突问题 foundation model
12 Beyond State Machines: Executing Network Procedures with Agentic Tool-Calling Sequences 利用Agentic Tool-Calling序列执行网络程序，提升移动通信系统灵活性。 large language model
13 Strategy-Aware Optimization Modeling with Reasoning LLMs 提出SAGE框架，显式建模优化策略，提升LLM在优化问题建模中的正确性和效率。 large language model ✅
14 From Experimental Limits to Physical Insight: A Retrieval-Augmented Multi-Agent Framework for Interpreting Searches Beyond the Standard Model 提出HEP-CoPilot，一个检索增强的多Agent框架，用于解释超出标准模型的搜索结果。 multimodal
15 GRAIL: A Deep-Granularity Hybrid Resonance Framework for Real-Time Agent Discovery via SLM-Enhanced Indexing GRAIL：通过SLM增强索引实现实时Agent发现的深度粒度混合共振框架 large language model
16 LLM-Assisted Repository-Level Generation with Structured Spec-Driven Engineering 提出结构化规约驱动工程(SSDE)，提升LLM在仓库级代码生成的质量和可验证性。 large language model
17 APIOT: Autonomous Vulnerability Management Across Bare-Metal Industrial OT Networks APIOT：实现裸机工业OT网络漏洞自主管理的框架 large language model
18 LLM-enabled Social Agents 提出基于角色定义的LLM社会智能体框架，提升社会交互能力 large language model
19 EngiAgent: Fully Connected Coordination of LLM Agents for Solving Open-ended Engineering Problems with Feasible Solutions EngiAgent：全连接LLM智能体协同解决可行性导向的开放式工程问题 large language model ✅
20 Complexity Horizons of Compressed Models in Analog Circuit Analysis 提出基于前提图的模型压缩策略，优化LLM在电路分析中的推理效率。 large language model ✅
21 On the Privacy of LLMs: An Ablation Study 针对LLM隐私风险，提出统一威胁模型并进行消融研究，揭示设计选择的影响。 large language model
22 Zero-Shot Confidence Estimation for Small LLMs: When Supervised Baselines Aren't Worth Training 针对小LLM，提出零样本置信度估计方法，无需监督训练即可实现可靠的本地-云路由。 large language model
23 CoVSpec: Efficient Device-Edge Co-Inference for Vision-Language Models via Speculative Decoding 提出CoVSpec，通过推测解码实现视觉-语言模型在端-边协同推理中的高效部署。 multimodal
24 Retrieval and Multi-Hop Reasoning in 1M-Token Context Windows: Evaluating LLMs on Classical Chinese Text 评估百万Token上下文窗口下LLM在古文检索与多跳推理能力 large language model
25 DocSync: Agentic Documentation Maintenance via Critic-Guided Reflexion DocSync：提出一种基于评论家引导反思的Agent，用于维护软件文档与代码的一致性。 large language model
26 The Dynamic Gist-Based Memory Model (DGMM): A Memory-Centric Architecture for Artificial Intelligence 提出动态概要记忆模型(DGMM)，解决AI在持久记忆、时序定位和可解释性方面的局限。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
27 Shadow-Loom: Causal Reasoning over Graphical World Model of Narratives Shadow-Loom：构建叙事图世界模型，实现因果推理与叙事物理分析 world model world models large language model
28 Standing on the Shoulders of Giants: Stabilized Knowledge Distillation for Cross--Language Code Clone Detection 提出基于稳定知识蒸馏的跨语言代码克隆检测方法，提升小型开源模型的可靠性。 distillation large language model
29 The 2026 ACII Dyadic Conversations (DaiKon) Workshop & Challenge ACII-DaiKon：用于建模对话情感与社交动态的双人对话基准挑战赛 MAE dyadic interaction multimodal
30 T$^2$PO: Uncertainty-Guided Exploration Control for Stable Multi-Turn Agentic Reinforcement Learning 提出T$^2$PO，通过不确定性引导探索控制，提升多轮Agent强化学习的稳定性。 reinforcement learning dreamer ✅
31 Reinforcement Learning Trained Observer Control for Bearings-Only Tracking 提出深度强化学习控制策略以解决目标跟踪问题 reinforcement learning deep reinforcement learning

⬅️ 返回 cs.AI 首页 · 🏠 返回主页