cs.AI（2026-04-23）

📊 共 32 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (25 🔗5) 支柱二：RL算法与架构 (RL & Architecture) (5) 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) (25 篇)

#题目一句话要点标签🔗⭐
1 Symbolic Grounding Reveals Representational Bottlenecks in Abstract Visual Reasoning 提出符号基础方法以解决抽象视觉推理瓶颈问题 large language model multimodal visual grounding
2 Learning to Communicate: Toward End-to-End Optimization of Multi-Agent Language Systems 提出DiffMAS框架，端到端优化多智能体语言系统的隐式通信。 large language model
3 GS-Quant: Granular Semantic and Generative Structural Quantization for Knowledge Graph Completion GS-Quant：用于知识图谱补全的粒度语义和生成结构量化方法 large language model ✅
4 How English Print Media Frames Human-Elephant Conflicts in India 利用Web规模文本分析揭示印度英语媒体中人象冲突的负面框架 large language model
5 Ideological Bias in LLMs' Economic Causal Reasoning 揭示LLM在经济因果推理中存在的意识形态偏见，尤其是在干预导向与市场导向观点对立时。 large language model
6 Spatial Metaphors for LLM Memory: A Critical Analysis of the MemPalace Architecture MemPalace：一种基于空间隐喻的LLM记忆系统，虽有夸大但具架构洞见 large language model
7 Can MLLMs "Read" What is Missing? 提出MMTR-Bench基准，评估多模态大语言模型从视觉上下文重建文本的能力 large language model multimodal instruction following ✅
8 ReCAPA: Hierarchical Predictive Correction to Mitigate Cascading Failures 提出ReCAPA框架，通过分层预测校正缓解视觉-语言-动作系统中级联失效问题 vision-language-action VLA large language model
9 Foundation models for discovering robust biomarkers of neurological disorders from dynamic functional connectivity 提出RE-CONFIRM框架与Hub-LoRA微调方法，提升脑疾病生物标志物识别的鲁棒性。 foundation model ✅
10 Transient Turn Injection: Exposing Stateless Multi-Turn Vulnerabilities in Large Language Models 提出瞬态轮次注入攻击，揭示大语言模型中无状态多轮对话漏洞 large language model
11 Enhancing Online Recruitment with Category-Aware MoE and LLM-based Data Augmentation 提出基于LLM数据增强和类别感知MoE的在线招聘方法，提升人岗匹配效果。 large language model chain-of-thought
12 CoFEE: Reasoning Control for LLM-Based Feature Discovery CoFEE：基于LLM的特征发现推理控制框架，提升特征质量与效率 large language model
13 Emergent Strategic Reasoning Risks in AI: A Taxonomy-Driven Evaluation Framework 提出ESRRSim框架，评估大型语言模型中涌现的战略推理风险 large language model
14 Ethics Testing: Proactive Identification of Generative AI System Harms 提出伦理测试，主动识别生成式AI系统中的潜在危害 large language model
15 Memanto: Typed Semantic Memory with Information-Theoretic Retrieval for Long-Horizon Agents Memanto：面向长时程Agent的类型化语义记忆与信息论检索 large language model
16 Call-Chain-Aware LLM-Based Test Generation for Java Projects CAT：一种调用链感知的LLM测试生成方法，提升Java项目测试覆盖率。 large language model
17 Tool Attention Is All You Need: Dynamic Tool Gating and Lazy Schema Loading for Eliminating the MCP/Tools Tax in Scalable Agentic Workflows 提出Tool Attention机制，通过动态工具门控和延迟模式加载，消除可扩展Agent工作流中的MCP/Tools Tax。 large language model ✅
18 Thinking with Reasoning Skills: Fewer Tokens, More Accuracy 提出可重用推理技能以提高推理准确性和效率 chain-of-thought
19 DryRUN: On the Role of Public Tests in LLM-Driven Code Generation DryRUN：无需公共测试用例，LLM自主生成代码并纠错 large language model
20 Unbiased Prevalence Estimation with Multicalibrated LLMs 提出基于多重校准的大语言模型，解决类别流行度估计中的偏差问题 large language model
21 Efficient Agent Evaluation via Diversity-Guided User Simulation 提出DIVERT，通过多样性引导的用户模拟高效评估LLM客服Agent large language model
22 SparKV: Overhead-Aware KV Cache Loading for Efficient On-Device LLM Inference SparKV：面向设备端LLM推理的开销感知KV缓存加载框架 large language model
23 SQLyzr: A Comprehensive Benchmark and Evaluation Platform for Text-to-SQL SQLyzr：一个全面的Text-to-SQL基准测试与评估平台 large language model ✅
24 Trust but Verify: Introducing DAVinCI -- A Framework for Dual Attribution and Verification in Claim Inference for Language Models DAVinCI框架通过双重归因与验证提升语言模型生成声明的事实可靠性。 large language model
25 Doubly Saturated Ramsey Graphs: A Case Study in Computer-Assisted Mathematical Discovery 结合SAT求解器与LLM生成代码，发现双重饱和Ramsey图的无限族 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
26 ReaGeo: Reasoning-Enhanced End-to-End Geocoding with LLMs ReaGeo：提出基于LLM的端到端地理编码框架，增强空间推理能力。 reinforcement learning spatial relationship large language model
27 Align Generative Artificial Intelligence with Human Preferences: A Novel Large Language Model Fine-Tuning Method for Online Review Management 提出一种新颖的偏好微调方法以解决在线评论管理问题 curriculum learning large language model
28 MambaCSP: Hybrid-Attention State Space Models for Hardware-Efficient Channel State Prediction 提出MambaCSP，一种硬件高效的混合注意力状态空间信道状态预测模型 Mamba SSM state space model
29 Nemobot Games: Crafting Strategic AI Gaming Agents for Interactive Learning with Large Language Models Nemobot Games：利用大语言模型构建交互式学习的战略AI游戏智能体 reinforcement learning imitation learning large language model
30 Pre-trained LLMs Meet Sequential Recommenders: Efficient User-Centric Knowledge Distillation 提出一种高效的用户中心知识蒸馏方法，将预训练LLM知识融入序列推荐系统。 distillation large language model

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
31 Model Predictive Control of Hybrid Dynamical Systems 针对混合动态系统，提出基于混合时间域的MPC算法并验证其稳定性 MPC model predictive control

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

#题目一句话要点标签🔗⭐
32 HiCrew: Hierarchical Reasoning for Long-Form Video Understanding via Question-Aware Multi-Agent Collaboration 提出HiCrew，通过问题感知的多智能体协作进行长视频分层推理理解 spatiotemporal

⬅️ 返回 cs.AI 首页 · 🏠 返回主页