cs.AI（2025-12-21）

📊 共 21 篇论文 | 🔗 1 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (12) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗1) 支柱一：机器人控制 (Robot Control) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (12 篇)

#	题目	一句话要点	标签	🔗	⭐
1	Gabliteration: Adaptive Multi-Directional Neural Weight Modification for Selective Behavioral Alteration in Large Language Models	提出Gabliteration，通过自适应多方向权重修改实现大语言模型行为的精准选择性改变。	large language model
2	Multimodal Bayesian Network for Robust Assessment of Casualties in Autonomous Triage	提出一种多模态贝叶斯网络，用于自主分诊中对伤员的稳健评估。	multimodal
3	HARBOR: Holistic Adaptive Risk assessment model for BehaviORal healthcare	提出HARBOR，用于行为健康风险评估的自适应语言模型	large language model multimodal
4	Reflective Confidence: Correcting Reasoning Flaws via Online Self-Correction	提出反思置信度框架，通过在线自纠正提升LLM推理能力	large language model chain-of-thought
5	Beyond the Prompt: An Empirical Study of Cursor Rules	大规模实证研究揭示了光标规则在软件工程中项目上下文编码的关键作用	large language model
6	A Modular Reference Architecture for MCP-Servers Enabling Agentic BIM Interaction	提出模块化参考架构以解决BIM交互中的工具调用问题	large language model
7	Code2Doc: A Quality-First Curated Dataset for Code Documentation	Code2Doc：高质量代码文档生成数据集，解决现有数据集质量问题。	large language model
8	Explainable and Fine-Grained Safeguarding of LLM Multi-Agent Systems via Bi-Level Graph Anomaly Detection	提出XG-Guard，通过双层图异常检测实现LLM多智能体系统的可解释和细粒度安全防护。	large language model
9	Social Comparison without Explicit Inference of Others' Reward Values: A Constructive Approach Using a Probabilistic Generative Model	利用概率生成模型，研究猴子在无显式奖励推断下的社会比较机制	multimodal
10	Remoe: Towards Efficient and Low-Cost MoE Inference in Serverless Computing	Remoe：面向Serverless计算的高效低成本MoE推理系统	large language model
11	AI Code in the Wild: Measuring Security Risks and Ecosystem Shifts of AI-Generated Code in Modern Software	首个大规模实证研究揭示AI生成代码在软件生态中的安全风险与演变趋势	large language model
12	Multi-Agent LLM Committees for Autonomous Software Beta Testing	提出多代理LLM委员会框架以解决软件测试效率低下问题	large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (7 篇)

#	题目	一句话要点	标签	🔗	⭐
13	ESearch-R1: Learning Cost-Aware MLLM Agents for Interactive Embodied Search via Reinforcement Learning	提出ESearch-R1，通过强化学习优化成本感知的MLLM智能体，用于交互式具身搜索。	reinforcement learning PPO large language model
14	Structural Reinforcement Learning for Heterogeneous Agent Macroeconomics	提出结构化强化学习(SRL)方法，高效求解异质Agent宏观经济模型	reinforcement learning
15	CORE: Concept-Oriented Reinforcement for Bridging the Definition-Application Gap in Mathematical Reasoning	提出CORE框架以解决数学推理中的定义与应用差距问题	reinforcement learning large language model
16	Toward Training Superintelligent Software Agents through Self-Play SWE-RL	提出Self-play SWE-RL，通过自博弈强化学习训练超智能软件Agent	reinforcement learning large language model
17	A Multi-agent Text2SQL Framework using Small Language Models and Execution Feedback	提出MATS框架，利用小语言模型和执行反馈解决Text2SQL任务，性能媲美大型语言模型。	reinforcement learning large language model	✅
18	Vox Deorum: A Hybrid LLM Architecture for 4X / Grand Strategy Game AI -- Lessons from Civilization V	提出Vox Deorum混合架构，赋能LLM在4X游戏中进行宏观策略推理。	reinforcement learning large language model
19	Adaptive Accountability in Networked MAS: Tracing and Mitigating Emergent Norms at Scale	提出自适应责任框架，用于大规模网络化多智能体系统中涌现规范的追踪与缓解。	PPO reward shaping

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#	题目	一句话要点	标签	🔗	⭐
20	ChronoDreamer: Action-Conditioned World Model as an Online Simulator for Robotic Planning	ChronoDreamer：用于机器人规划的动作条件世界模型，作为在线模拟器	manipulation world model dreamer
21	Assignment-Routing Optimization: Solvers for Problems Under Constraints	提出基于MIP的联合路由-分配优化求解器，解决约束下的包装规划问题	manipulation mobile manipulation motion planning

⬅️ 返回 cs.AI 首页 · 🏠 返回主页