cs.AI（2026-02-24）

📊 共 27 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) (18 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (6) 支柱一：机器人控制 (Robot Control) (3)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (18 篇)

#题目一句话要点标签🔗⭐
1 NoRD: A Data-Efficient Vision-Language-Action Model that Drives without Reasoning 提出NoRD以解决数据收集与推理标注成本高的问题 vision-language-action VLA
2 Predicting Sentence Acceptability Judgments in Multimodal Contexts 研究视觉上下文对人类和LLM句子可接受性判断的影响 large language model multimodal
3 Physics-based phenomenological characterization of cross-modal bias in multimodal models 提出基于物理的表征方法，分析多模态大语言模型中的跨模态偏差问题。 large language model multimodal
4 Multimodal MRI Report Findings Supervised Brain Lesion Segmentation with Substructures 提出MS-RSuper方法，利用多模态MRI报告监督脑部病灶及其子结构的分割。 multimodal
5 E-MMKGR: A Unified Multimodal Knowledge Graph Framework for E-commerce Applications 提出E-MMKGR：一个用于电商应用的统一多模态知识图谱框架 multimodal
6 Qwen-BIM: developing large language model for BIM-based design with domain-specific benchmark and dataset Qwen-BIM：构建领域特定大语言模型，用于BIM设计，并提出相应基准和数据集。 large language model
7 Modality-Guided Mixture of Graph Experts with Entropy-Triggered Routing for Multimodal Recommendation 提出MAGNET模型，通过模态引导的图专家混合网络和熵触发路由，提升多模态推荐效果。 multimodal
8 Counterfactual Simulation Training for Chain-of-Thought Faithfulness 提出反事实模拟训练(CST)以提升思维链(CoT)推理的可靠性。 chain-of-thought ✅
9 PromptCD: Test-Time Behavior Enhancement via Polarity-Prompt Contrastive Decoding PromptCD：极性提示对比解码，提升LLM/VLM测试时行为可控性 large language model visual grounding
10 A Benchmark for Deep Information Synthesis 提出DEEPSYNTH基准，评估LLM智能体在复杂信息合成与推理任务中的能力。 large language model
11 SparkMe: Adaptive Semi-Structured Interviewing for Qualitative Insight Discovery SparkMe：自适应半结构化访谈，利用多智能体LLM进行定性洞察发现 large language model ✅
12 "Are You Sure?": An Empirical Study of Human Perception Vulnerability in LLM-Driven Agentic Systems 首个大规模人类实验揭示LLM驱动Agent系统中Agent介导欺骗的脆弱性 large language model
13 LogicGraph : Benchmarking Multi-Path Logical Reasoning via Neuro-Symbolic Generation and Verification LogicGraph：提出神经符号生成与验证框架，用于评估多路径逻辑推理能力。 large language model ✅
14 Architecting AgentOS: From Token-Level Context to Emergent System-Level Intelligence 提出AgentOS框架，将LLM重定义为推理内核，提升系统级智能 large language model
15 HELP: HyperNode Expansion and Logical Path-Guided Evidence Localization for Accurate and Efficient GraphRAG 提出HELP框架，通过超节点扩展和逻辑路径引导，提升GraphRAG的准确性和效率 large language model
16 AdapTools: Adaptive Tool-based Indirect Prompt Injection Attacks on Agentic LLMs AdapTools：针对Agentic LLM的自适应工具型间接提示注入攻击 large language model
17 Grounding LLMs in Scientific Discovery via Embodied Actions EmbodiedAct：通过具身动作将LLM应用于科学发现，解决长时程模拟中的可靠性和稳定性问题。 large language model
18 Hybrid LLM-Embedded Dialogue Agents for Learner Reflection: Designing Responsive and Theory-Driven Interactions 提出混合LLM嵌入式对话Agent，用于支持学习者反思，设计响应式和理论驱动的交互 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (6 篇)

#题目一句话要点标签🔗⭐
19 CG-DMER: Hybrid Contrastive-Generative Framework for Disentangled Multimodal ECG Representation Learning 提出CG-DMER框架，用于解耦多模态心电图表征学习，提升心血管疾病诊断准确性。 representation learning multimodal
20 Buffer Matters: Unleashing the Power of Off-Policy Reinforcement Learning in Large Language Model Reasoning 提出BAPO：一种基于Off-Policy RLVR的大语言模型推理能力提升方法 reinforcement learning large language model
21 Balancing Multiple Objectives in Urban Traffic Control with Reinforcement Learning from AI Feedback 提出多目标强化学习框架以优化城市交通控制 reinforcement learning policy learning reward design
22 OptiLeak: Efficient Prompt Reconstruction via Reinforcement Learning in Multi-tenant LLM Services OptiLeak：利用强化学习高效重建多租户LLM服务中的泄露提示 reinforcement learning direct preference optimization
23 PyVision-RL: Forging Open Agentic Vision Models via RL PyVision-RL：通过强化学习赋能开放Agentic视觉模型 reinforcement learning multimodal
24 PRECTR-V2:Unified Relevance-CTR Framework with Cross-User Preference Mining, Exposure Bias Correction, and LLM-Distilled Encoder Optimization PRECTR-V2：融合用户偏好挖掘、偏差校正和LLM蒸馏的统一相关性-CTR框架 representation learning distillation

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
25 Recursive Belief Vision Language Model 提出RB-VLA，通过置信度建模解决VLA模型在长时程操作中的部分可观测性问题 manipulation diffusion policy world model
26 Inner Speech as Behavior Guides: Steerable Imitation of Diverse Behaviors for Human-AI coordination MIMIC：利用内部语言指导，实现人机协作中行为的多样性模仿与可控引导 manipulation imitation learning behavior cloning
27 Pressure Reveals Character: Behavioural Alignment Evaluation at Depth 提出压力测试基准，揭示语言模型在复杂情境下的对齐问题 manipulation

⬅️ 返回 cs.AI 首页 · 🏠 返回主页