cs.AI（2026-07-02）

📊 共 38 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) (22 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (11) 支柱四：生成式动作 (Generative Motion) (3) 支柱一：机器人控制 (Robot Control) (1) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (22 篇)

#题目一句话要点标签🔗⭐
1 Multimodal Knowledge Edit-Scoped Generalization for Online Recursive MLLM Editing 提出ScopeEdit以解决多模态大语言模型在线编辑的语义边界问题 large language model multimodal ✅
2 Hidden Forgetting in Continual Multimodal Learning: When Accuracy Survives but Grounding Fails 提出RCL框架以解决多模态学习中的隐藏遗忘问题 large language model multimodal
3 InduceKV: Fixed-Footprint Continual Adaptation of Multimodal LLMs via Inducing KV Memories 提出InduceKV以解决多模态LLMs的持续适应问题 large language model multimodal
4 MMIR-TCM: Memory-Integrated Multimodal Inference and Retrieval for TCM Clinical Decision Support 提出MMIR-TCM以解决中医临床决策支持中的主观性与可重复性问题 large language model multimodal
5 OntoLearner: A Modular Python Library for Ontology Learning with Large Language Models 提出OntoLearner以解决本体学习领域的基础设施不足问题 large language model ✅
6 Pre-Flight: A Benchmark for Evaluating Large Language Models on Aviation Operational Knowledge 提出Pre-Flight基准以评估航空领域大型语言模型的操作知识 large language model
7 Fast Multi-dimensional Refusal Subspaces via RFM-AGOP 提出RFM-AGOP以快速识别多维拒绝子空间 large language model
8 AgenticDataBench: A Comprehensive Benchmark for Data Agents 提出AgenticDataBench以解决数据代理评估不足问题 large language model
9 ReContext: Recursive Evidence Replay as LLM Harness for Long-Context Reasoning 提出RECONTEXT以解决长上下文推理中的证据利用问题 large language model ✅
10 TestEvo-Bench: An Executable and Live Benchmark for Test and Code Co-Evolution 提出TestEvo-Bench以解决测试与代码共同演化问题 foundation model
11 SkillFuzz: Fuzzing Skill Composition for Implicit Intents Discovery in Open Skill Marketplaces 提出SkillFuzz以解决隐式意图发现问题 large language model
12 SelectTSL: Prompt-Guided Selective Target Sound Localization in Complex Scenarios 提出SelectTSL以解决复杂场景中的目标声音定位问题 multimodal
13 UA-ChatDev: Uncertainty-Aware Multi-Agent Collaboration for Reliable Software Development 提出UA-ChatDev以解决软件开发中的不确定性问题 large language model
14 Behind the Refusal: Determining Guardrail Activation via Behavioral Monitoring 提出黑箱守卫侦察方法以解决AI安全监测问题 large language model
15 Prompt Coverage Adequacy 提出Prompt Coverage Adequacy以解决LLM驱动软件开发测试问题 large language model
16 Robust for the Wrong Reasons: The Representational Geometry of LLM Robustness to Science Skepticism 提出对大型语言模型在科学怀疑下鲁棒性的深入分析 large language model
17 ContextSniper: AntTrail's Token-Efficient Code Memory for Repository-Level Program Repair 提出ContextSniper以解决代码修复中的上下文效率问题 large language model ✅
18 Safety Targeted Embedding Exploit via Refinement 提出STEER以解决多语言安全机制泛化不足问题 large language model
19 SimWorlds: A Multi-Agent System for Dynamic 3D Scene Creation 提出SimWorlds以解决动态3D场景生成问题 embodied AI
20 Path-level Hindsight Instructions for Semantic Exploration in Vision-Language Navigation 提出Phi-Nav以解决视觉语言导航中的语义探索问题 VLN
21 ADVENT: LLM-Driven Automatic Predicate Invention for ILP 提出ADVENT以解决归纳逻辑编程中的谓词发明问题 large language model
22 EO-Agents: A Three-Agent LLM Pipeline for Earth Observation Hypothesis Generation 提出EO-Agents以解决地球观测假设生成问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (11 篇)

#题目一句话要点标签🔗⭐
23 Safe and Adaptive Cloud Healing: Verifying LLM-Generated Recovery Plans with a Neural-Symbolic World Model 提出PASE框架以解决云服务自愈问题 reinforcement learning deep reinforcement learning DRL
24 Enhancing Fitness Intelligence through Domain-Specific LLM Post-Training 提出FitOne以解决科学健身教练领域的知识不足问题 reinforcement learning large language model foundation model
25 Repair the Amplifier, Not the Symptom: Stable World-Model Correction for Agent Rollouts 提出WM-SAR方法以解决大规模规划图中的错误修复问题 world model world models
26 DRL-CLBA: A Clean Label Backdoor Attack for Speech Classification via DDPG Reinforcement Learning 提出DRL-CLBA以解决语音分类中的清标签后门攻击问题 reinforcement learning DRL
27 Scaling with Confidence: Calibrating Confidence of LLMs for Adaptive Test Time Scaling 提出C3RL与CAS以解决LLMs信心校准问题 reinforcement learning reward design large language model
28 Purified OPSD: On-Policy Self-Distillation Without Losing How to Think 提出Purified OPSD以解决长链推理模型的自我蒸馏问题 distillation chain-of-thought
29 Hardware-Enforced Semantic Coordination for Safety-Critical Real-Time Autonomous Systems 提出硬件强制语义协调架构以解决安全关键实时系统的协调问题 world model world models large language model
30 DRIFTLENS: Measuring Memory-Induced Reasoning Drift in Personalized Language Models 提出DRIFTLENS框架以测量个性化语言模型中的推理漂移 DPO instruction following
31 A Hippocampus for Linear Attention: An Exact Memory for What the Recurrent State Forgets 提出HOLA以解决线性注意力模型的记忆遗忘问题 linear attention
32 Hawk: Harnessing Hardware-Aware Knowledge for High-Performance NPU Kernel Generation 提出Hawk框架以解决NPU内核生成中的硬件约束问题 distillation large language model
33 Scaling Trends for Lie Detector Oversight in Preference Learning 提出可扩展的谎言检测器监督方法以解决偏好学习中的欺骗行为监测问题 preference learning

🔬 支柱四：生成式动作 (Generative Motion) (3 篇)

#题目一句话要点标签🔗⭐
34 Text-Driven 3D Indoor Scene Synthesis in Non-Manhattan Environments 提出SPG-Layout以解决非曼哈顿环境中的室内场景合成问题 physically plausible spatial relationship large language model
35 An Efficient vLLM-Based Inference Pipeline for Unified Audio Understanding and Generation 提出基于vLLM的推理管道以解决多模态音频理解与生成问题 classifier-free guidance multimodal
36 Low-Latency Task-Oriented Image Transmission with Opportunistic Spectrum Access 提出低延迟任务导向图像传输框架以解决频谱有限问题 VQ-VAE

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
37 Automated grading of Linux/bash examinations using large language models: a four-level cognitive taxonomy approach 利用大型语言模型实现Linux/bash考试的自动评分 manipulation MAE large language model

🔬 支柱八：物理动画 (Physics-based Animation) (1 篇)

#题目一句话要点标签🔗⭐
38 Scene-Conditioned PINN-GNN for Multipath RF Maps: Cross-Scene Generation and In-Scene Completion 提出基于PINN-GNN的框架以解决多路径RF地图生成问题 PULSE

⬅️ 返回 cs.AI 首页 · 🏠 返回主页