cs.AI（2026-05-15）

📊 共 29 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (16 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (10 🔗1) 支柱一：机器人控制 (Robot Control) (1 🔗1) 支柱七：动作重定向 (Motion Retargeting) (1) 支柱三：空间感知与语义 (Perception & Semantics) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (16 篇)

#题目一句话要点标签🔗⭐
1 Towards Foundation Models for Relational Databases with Language Models and Graph Neural Networks 提出结合语言模型和图神经网络的关系数据库Foundation模型 foundation model
2 SaaS-Bench: Can Computer-Use Agents Leverage Real-World SaaS to Solve Professional Workflows? SaaS-Bench：评估计算机使用Agent在真实SaaS环境中解决专业工作流的能力 large language model multimodal ✅
3 DRS-GUI: Dynamic Region Search for Training-Free GUI Grounding DRS-GUI：免训练动态区域搜索，提升GUI界面元素定位精度 large language model multimodal
4 See Before You Code: Learning Visual Priors for Spatially Aware Educational Animation Generation OmniManim：基于视觉先验的空间感知教育动画生成框架 large language model
5 Context, Reasoning, and Hierarchy: A Cost-Performance Study of Compound LLM Agent Design in an Adversarial POMDP 在对抗性POMDP中，研究复合LLM Agent设计的成本效益，并提出优化策略。 chain-of-thought
6 PrismQuant: Rate-Distortion-Optimal Vector Quantization for Gaussian-Mixture Sources PrismQuant：针对高斯混合源的率失真最优矢量量化方法 multimodal
7 Prospective multi-pathogen disease forecasting using autonomous LLM-guided tree search 提出基于LLM引导树搜索的自主多病原体疾病预测系统，克服人工建模瓶颈。 large language model
8 Reasoners or Translators? Contamination-aware Evaluation and Neuro-Symbolic Robustness in Tax Law 提出污染感知评估方法，并验证神经符号框架在税法推理中更具鲁棒性和泛化性。 large language model
9 Toward Natural and Companionable Virtual Agents via Cross-Temporal Emotional Modeling 提出跨时间情感建模框架CTEM，提升虚拟陪伴型Agent的自然性和连贯性 foundation model
10 Can We Trust AI-Inferred User States. A Psychometric Framework for Validating the Reliability of Users States Classification by LLMs in Operational Environments 提出评估框架，验证LLM推断用户状态的可靠性，提升自适应系统AI设计的可信度。 large language model
11 Position: Early-Stage Quality Assurance in Annotation Pipelines Is More Cost-Effective Than Late-Stage Validation 在标注流程中，早期质量保证比后期验证更具成本效益 foundation model
12 ColPackAgent: Agent-Skill-Guided Hard-Particle Monte Carlo Workflows for Colloidal Packing 提出ColPackAgent，通过Agent-Skill引导的硬粒子蒙特卡洛工作流进行胶体堆积模拟 large language model
13 A Few GPUs, A Whole Lotta Scale: Faithful LLM Training Emulation with PrismLLM PrismLLM：利用少量GPU实现大规模LLM训练的忠实仿真 large language model
14 Detecting Privilege Escalation in Polyglot Microservices via Agentic Program Analysis Neo：利用Agentic程序分析检测Polyglot微服务中的权限提升漏洞 large language model
15 RTL-BenchMT: Dynamic Maintenance of RTL Generation Benchmark Through Agent-Assisted Analysis and Revision 提出RTL-BenchMT框架，利用智能体辅助动态维护RTL生成基准测试集。 large language model
16 CAPS: Cascaded Adaptive Pairwise Selection for Efficient Parallel Reasoning 提出CAPS：级联自适应配对选择，用于高效并行推理 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (10 篇)

#题目一句话要点标签🔗⭐
17 Imperfect World Models are Exploitable 提出模型利用新定义，揭示强化学习中不完善世界模型的潜在风险。 reinforcement learning world model world models
18 Deterministic Event-Graph Substrates as World Models for Counterfactual Reasoning 提出基于确定性事件图基质的世界模型，用于反事实推理。 world model world models
19 Structure Abstraction and Generalization in a Hippocampal-Entorhinal Inspired World Model 提出一种受海马-内嗅皮层启发的结构抽象世界模型，实现结构泛化 world model world models
20 Nudging Beyond the Comfort Zone: Efficient Strategy-Guided Exploration for RLVR NudgeRL：基于策略引导的高效探索RLVR框架，提升LLM推理能力 reinforcement learning distillation privileged information ✅
21 Look Before You Leap: Autonomous Exploration for LLM Agents 提出Explore-then-Act范式，提升LLM Agent在未知环境下的自主探索能力 reinforcement learning affordance large language model
22 ALSO: Adversarial Online Strategy Optimization for Social Agents 提出ALSO框架，通过对抗在线策略优化提升社交智能体在动态环境中的适应性。 reinforcement learning offline reinforcement learning large language model
23 PAGER: Bridging the Semantic-Execution Gap in Point-Precise Geometric GUI Control PAGER：弥合点精确几何GUI控制中的语义-执行鸿沟 reinforcement learning multimodal
24 Agentic Discovery of Neural Architectures: AIRA-Compose and AIRA-Design 提出AIRA框架，利用LLM自主设计超越Transformer的下一代基础模型 Mamba foundation model
25 Access Timing as Scaffolding: A Reinforcement Learning Approach to GenAI in Education 提出基于强化学习的GenAI访问时机控制方法，提升教育场景下的学习效果和元认知参与度。 reinforcement learning
26 TopoEvo: A Topology-Aware Self-Evolving Multi-Agent Framework for Root Cause Analysis in Microservices 提出TopoEvo框架以解决微服务中的根因分析问题 representation learning multimodal

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
27 Learning Bilevel Policies over Symbolic World Models for Long-Horizon Planning 提出BISON，通过符号世界模型学习双层策略，解决长时程规划问题。 manipulation imitation learning world model ✅

🔬 支柱七：动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗⭐
28 Symplectic Neural Operators for Learning Infinite Dimensional Hamiltonian Systems 提出辛神经网络算子，用于学习无限维哈密顿系统，保证长期稳定性。 structure preservation

🔬 支柱三：空间感知与语义 (Perception & Semantics) (1 篇)

#题目一句话要点标签🔗⭐
29 ShopGym: An Integrated Framework for Realistic Simulation and Scalable Benchmarking of E-Commerce Web Agents ShopGym：用于电商Web Agent的逼真模拟与可扩展基准测试的集成框架 affordance

⬅️ 返回 cs.AI 首页 · 🏠 返回主页