cs.AI（2026-08-28）

📊 共 38 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) (24 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (9) 支柱一：机器人控制 (Robot Control) (3) 支柱三：空间感知与语义 (Perception & Semantics) (1) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (24 篇)

#题目一句话要点标签🔗⭐
1 RetailAgent: Structured Adverse Timing in Self-Conditioned Multimodal LLM Trading Agents 提出RetailAgent框架以解决金融市场中的可预测性问题 large language model multimodal
2 CoRe-MoE: Compact Reusable MoE for Continual Multimodal Instruction Tuning 提出CoRe-MoE以解决持续多模态指令调优中的参数冗余问题 large language model multimodal ✅
3 When Robots Mishear Us: Mapping the Safety Risks of Voice-Controlled Embodied AI 研究ASR错误对具身AI安全风险的影响 embodied AI
4 Do Medical Vision Models Reason About Anatomy? Probing the Spatial Inductive Biases of Learned Visual Representations 构建SPAR-Bench以探讨医学视觉模型的解剖推理能力 foundation model multimodal zero-shot transfer
5 MAP: A Benchmark on Multimodal Accessibility Planning for Real World Places 提出MAP基准以解决多模态无障碍规划问题 multimodal
6 WeAgent-MMSearch: Native Text-Vision Interaction for Multimodal Search Agents 提出WeAgent-MMSearch以解决多模态搜索代理中的视觉信息缺失问题 multimodal
7 Automated Analysis Framework for Multilingual Climate-Health Literature Based on Multi-Agent Large Language Model 提出多代理大语言模型框架以解决气候健康文献分析问题 large language model
8 See, Hypothesize, Validate: Multimodal Agentic Framework for Discovering Governing PDEs 提出MAGE框架以解决偏微分方程发现问题 multimodal
9 The Shape of Power: A Multilingual Framework for Social Power Reasoning in Dialogues 提出多语言框架以解决社会权力推理的跨文化分析问题 large language model multimodal
10 Compared to What? A Human-Anchored Security Benchmark for LLM-Generated Infrastructure-as-Code 提出GenIaC-SecBench以评估LLM生成的基础设施代码安全性 large language model chain-of-thought
11 Learning a Size-Weight Frontier for Synthetic-Augmented Inference 提出合成增强推断的大小-权重前沿以解决数据稀缺问题 large language model
12 GRACE:Gradient-guided Coreset Selection for LLM Unlearning 提出GRACE以解决大语言模型的遗忘与保留数据选择问题 large language model
13 Stay Within Your Bounds: Distance-Guided Decoding for Guaranteed Context-Free Grammar Compliance 提出基于推送自动机的解码框架以确保上下文无关文法合规性 large language model
14 Not to Break, but to Attest: Adversarial Probes for Privacy-Preserving LLM Verification 提出基于zk-SNARK的审计框架以解决LLM隐私验证问题 large language model
15 LLM-Based Agents for Software and Systems Security: Approaches, Applications, and Assessment 提出基于大型语言模型的智能体以提升软件和系统安全性 large language model
16 LongPIBench: A Long-Context Benchmark for Prompt Injection 提出LongPIBench以解决长上下文中的提示注入攻击问题 large language model
17 Layered LLM Defenses as an Ensemble: Access Tiers, Inference Cost, and the Measured Failure Correlation Between Defense Layers 提出分层LLM防御体系以解决模型安全性问题 large language model
18 MAIL: Memory-driven, Adaptive, Incremental, and Literature-grounded Framework for Hypothesis Generation in Chemistry 提出MAIL框架以解决化学假设生成中的知识导航问题 large language model
19 Beyond Task-Only Matching: Personalized Skill Routing with Counterfactual Evaluation 提出个性化技能路由方法以解决任务匹配不足问题 large language model
20 REINS: Refusal-Enhanced Inhibitory Steering with Sparse Autoencoder Features 提出REINS以增强稀疏自编码器特征的拒绝能力 large language model
21 CAITLYN: Can LLM Agents Autonomously Synthesize Defenses against Emerging Injection Attacks? 提出CAITLYN以解决大语言模型的注入攻击防御问题 large language model
22 When Evidence Shapes Collaboration: Knowledge-Conditioned Topology Generation for Multi-Agent Systems 提出K-GAT以解决多智能体系统中的协作拓扑生成问题 large language model
23 AI Alignment through a Game-theoretic Lens: A Survey 通过博弈论视角提出AI对齐方法以应对复杂人类价值问题 large language model
24 FISGuard: Defending Against Membership Inference via Fixed Input Subspaces 提出FISGuard以解决成员推断攻击问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (9 篇)

#题目一句话要点标签🔗⭐
25 Learning to Allocate Incentives for Incentivized Advertising via Offline Model-Based Reinforcement Learning 提出基于离线模型强化学习的激励广告分配方法 reinforcement learning TD3 offline RL
26 AcrossVAM1.0: Particle World Modeling for Text-Assisted Robot Video Prediction 提出AcrossVAM1.0以解决机器人视频预测中的运动推理与外观保留问题 world model world models
27 Learning to Use Tools: Reinforcement Learning for Tool-Integrated Mathematical Reasoning 提出工具集成强化学习以提升数学推理能力 reinforcement learning large language model
28 When Teacher Guidance Misleads: Reward-Aligned On-Policy Distillation 提出奖励对齐的在线蒸馏方法以解决教师指导误导问题 distillation large language model
29 Physics-Guided Flow Matching for CT Image Reconstruction 提出基于物理引导的流匹配方法以解决CT图像重建问题 flow matching
30 Finding Where the Buck Stops: An Automated Failure Attribution-Based Reflection Framework for Multi-Agent Collaboration 提出DoCtOR框架以解决多智能体系统中的失败归因问题 PPO large language model
31 Rubric-to-Code Credit Assignment for Reinforcement Learning 提出Rubric-to-Code信用分配以解决交互式应用生成问题 reinforcement learning
32 SpikeOPD: Stable On-Policy Distillation for Autoregressive Spiking Language Models 提出SpikeOPD以解决自回归脉冲语言模型的稳定性问题 distillation
33 ReToolSQL: Agentic Reinforcement Learning for Robust Text-to-SQL 提出ReToolSQL以解决文本到SQL生成中的迭代错误问题 reinforcement learning

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
34 Low-Altitude Fluid Antenna Network with Multi-Agent Reinforcement Learning 提出基于多智能体强化学习的低空流体天线网络优化方案 sim-to-real reinforcement learning
35 Should I Use This Synthetic Dataset for Training? How to Test with Minimal Real Data 提出aeSFT以解决合成数据训练效果评估问题 sim-to-real world model world models
36 Generative AI Alignment with Hinduism's Theological Plurality and Sacred Representation 探讨生成性人工智能与印度教神学多元性及神圣表现的对齐问题 manipulation

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

#题目一句话要点标签🔗⭐
37 CURA: Certified Runtime Alarms for Computer-Use Agents 提出CURA以解决计算机使用代理的运行时监控问题 affordance

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

#题目一句话要点标签🔗⭐
38 Logos: An Agent Harness on a Cross-Process Bus 提出Logos以解决代理系统跨进程协作问题 spatiotemporal

⬅️ 返回 cs.AI 首页 · 🏠 返回主页