cs.AI（2026-08-26）

📊 共 22 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (15 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (3) 支柱一：机器人控制 (Robot Control) (2) 支柱八：物理动画 (Physics-based Animation) (1) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (15 篇)

#题目一句话要点标签🔗⭐
1 SciMIF: Understanding Multimodal Instruction Following in Scientific Domains 提出SciMIF以评估科学领域中的多模态指令遵循能力 large language model multimodal instruction following ✅
2 Multi-Granularity Context-Enhanced RAG over Multimodal Knowledge Graphs 提出多粒度上下文增强的MMKG以提升多模态RAG性能 large language model multimodal
3 MMJailBench: A Factorized Benchmark for Disentangling Multimodal Jailbreak Vulnerabilities 提出MMJailBench以解决多模态大型语言模型的监狱突破脆弱性问题 large language model multimodal
4 Finding and using interpretable latents in a neutrino foundation model with sparse autoencoders 提出稀疏自编码器以实现中微子模型的可解释性 foundation model
5 Data Citation for Large Language Models: A Challenge 提出数据引用机制以解决大型语言模型的引用挑战 large language model
6 Paint What You See: Benchmarking Dexterous Visual Tool Use in Multimodal Agents 提出EASEL基准以解决多模态智能体的精细视觉工具使用问题 multimodal
7 FinRiskAtlas: Decision-Aligned Evaluation of Large Language Models for Financial Risk Review 提出FinRiskAtlas以解决金融风险审查中的模型评估问题 large language model
8 AI Slop and Hallucinations in Vulnerability Assessment: A Survey on Reasoning Failures and Trustworthy Mitigation 提出主动神经符号验证以解决AI漏洞评估中的信任危机 large language model chain-of-thought
9 Narcissus: Program Synthesis Using Context-Aware LLM Approximations 提出Narcissus以解决编程语言固定任务的合成问题 large language model
10 CRAMER: Control via Request-Aware Masking for Editing Recommenders 提出CRAMER框架以解决推荐系统对用户请求响应不足的问题 large language model
11 Repair or Resample? Rethinking Failure Debugging in LLM Multi-Agent Systems 提出SymTrace框架以解决LLM多智能体系统调试问题 large language model
12 ToST: A Tree-of-Thought Socratic Teaching Framework for Multi-Path Guidance and Parallel Thinking 提出ToST框架以解决单路径教学限制问题 large language model
13 CaSKG: Counterfactual-Causal Skill Graphs for Scalable Agent Skill Retrieval 提出CaSKG框架以解决大语言模型技能检索问题 large language model ✅
14 MACGen: Toward Functionally Correct and Secure Code Generation via Multi-Agent Collaboration 提出MACGen以解决安全与功能正确性代码生成问题 large language model
15 Q&A or Document-Based? The Effects of Interface Type on How Screen Reader Users Access Interconnected Documents 比较问答界面与文档界面对盲人用户知识构建的影响 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (3 篇)

#题目一句话要点标签🔗⭐
16 Agentic Game Development as a Verifiable Trajectory Data Engine for Scaling World Models 提出RLHEV以解决世界模型奖励信号不足问题 reinforcement learning world model world models
17 Imitation Learning for Connection-Tableau Construction 提出模仿学习方法以构建连接表格 policy learning imitation learning
18 Training Alignment Auditors via Reinforcement Learning 通过强化学习提升大型语言模型审计员的对齐审计能力 reinforcement learning

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
19 A Training-Free Proactive Defense Against Partial Speech Manipulation via Self-Embedding Steganography 提出自嵌入隐写术以应对部分语音操控问题 manipulation
20 Combining Self-Embedding Audio Watermarking with Ultra-Low-Bitrate Neural Codecs 提出自嵌入音频水印与超低比特率神经编码器结合以解决内容完整性验证问题 manipulation

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

#题目一句话要点标签🔗⭐
21 Planetary Prediction Engine: Autonomous Geospatial Prediction via Intelligent Data Selection and Foundation Model Embeddings 提出行星预测引擎以解决全球地理预测问题 spatiotemporal foundation model multimodal

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
22 Unlocking Multimodal Protein Language Models at Inference Time 提出三阶段框架优化多模态蛋白质语言模型推理策略 classifier-free guidance multimodal

⬅️ 返回 cs.AI 首页 · 🏠 返回主页