cs.AI（2026-04-13）

📊 共 41 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) (27 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (12 🔗1) 支柱八：物理动画 (Physics-based Animation) (1) 支柱一：机器人控制 (Robot Control) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (27 篇)

#题目一句话要点标签🔗⭐
1 CFMS: A Coarse-to-Fine Multimodal Synthesis Framework for Enhanced Tabular Reasoning 提出CFMS框架以增强表格推理能力 large language model multimodal chain-of-thought
2 Dynamic Summary Generation for Interpretable Multimodal Depression Detection 提出基于大语言模型的多阶段框架，用于可解释的多模态抑郁症检测。 large language model multimodal
3 Environmental Footprint of GenAI Research: Insights from the Moshi Foundation Model 细粒度分析Moshi模型研发全流程，揭示并降低GenAI研究的环境足迹 large language model foundation model
4 EmergentBridge: Improving Zero-Shot Cross-Modal Transfer in Unified Multimodal Embedding Models 提出EmergentBridge以解决跨模态无监督对齐问题 multimodal zero-shot transfer
5 Anthropogenic Regional Adaptation in Multimodal Vision-Language Model 提出人类中心区域自适应范式，优化多模态视觉语言模型在特定区域的文化相关性。 multimodal
6 Back to the Barn with LLAMAs: Evolving Pretrained LLM Backbones in Finetuning Vision Language Models 研究LLM骨干演进对视觉语言模型的影响，揭示性能与任务依赖性 large language model multimodal instruction following
7 Why Do Large Language Models Generate Harmful Content? 提出基于因果中介分析的方法，探究大语言模型生成有害内容的原因。 large language model
8 Intersectional Sycophancy: How Perceived User Demographics Shape False Validation in Large Language Models 研究表明大型语言模型中的虚假肯定行为受用户人口统计特征影响 large language model
9 Consistency of AI-Generated Exercise Prescriptions: A Repeated Generation Study Using a Large Language Model 评估LLM生成运动处方的一致性：一项基于Gemini 2.5 Flash的重复生成研究 large language model
10 Delving Aleatoric Uncertainty in Medical Image Segmentation via Vision Foundation Models 利用视觉基础模型估计医学图像分割中的本征不确定性，提升模型鲁棒性 foundation model
11 Beyond A Fixed Seal: Adaptive Stealing Watermark in Large Language Models 提出自适应窃取水印算法，提升针对大语言模型水印的攻击效率。 large language model ✅
12 Agentic Driving Coach: Robustness and Determinism of Agentic AI-Powered Human-in-the-Loop Cyber-Physical Systems 提出基于反应模型的AI驱动教练以解决人机协作系统中的不确定性问题 large language model foundation model
13 Mobile GUI Agent Privacy Personalization with Trajectory Induced Preference Optimization 提出TIPO，通过轨迹诱导偏好优化实现移动GUI代理的隐私个性化 large language model multimodal ✅
14 Diffusion-CAM: Faithful Visual Explanations for dMLLMs 提出Diffusion-CAM，为扩散多模态大语言模型提供可靠的可视化解释。 large language model multimodal
15 Audio Flamingo Next: Next-Generation Open Audio-Language Models for Speech, Sound, and Music 提出Audio Flamingo Next，用于提升语音、声音和音乐理解的下一代开放音频语言模型。 chain-of-thought TAMP
16 The Salami Slicing Threat: Exploiting Cumulative Risks in LLM Systems 提出Salami Attack，利用累积风险突破LLM安全防线，实现多模态通用越狱 large language model
17 Min-$k$ Sampling: Decoupling Truncation from Temperature Scaling via Relative Logit Dynamics 提出Min-$k$采样方法，通过相对Logit动态解耦截断与温度缩放，提升大语言模型文本生成质量。 large language model
18 CASK: Core-Aware Selective KV Compression for Reasoning Traces CASK：面向推理轨迹的核心感知选择性KV压缩，提升长文本推理性能 large language model
19 ClawGuard: A Runtime Security Framework for Tool-Augmented LLM Agents Against Indirect Prompt Injection ClawGuard：针对工具增强型LLM Agent的运行时安全框架，防御间接Prompt注入攻击 large language model ✅
20 SWE-AGILE: A Software Agent Framework for Efficiently Managing Dynamic Reasoning Context SWE-AGILE：提出动态推理上下文管理的软件Agent框架，提升软件工程任务效率。 chain-of-thought ✅
21 DreamKG: A KG-Augmented Conversational System for People Experiencing Homelessness DreamKG：一个知识图谱增强的对话系统，服务于无家可归者 large language model
22 A collaborative agent with two lightweight synergistic models for autonomous crystal materials research MatBrain：轻量级协同智能体加速晶体材料自主研究 large language model
23 From Translation to Superset: Benchmark-Driven Evolution of a Production AI Agent from Rust to Python 提出基于基准测试驱动的LLM辅助代码迁移方法，实现Rust到Python的AI Agent演进 large language model
24 SLALOM: Simulation Lifecycle Analysis via Longitudinal Observation Metrics for Social Simulation SLALOM：通过纵向观察指标分析社会模拟生命周期，解决LLM社会模拟验证难题 large language model
25 Network Effects and Agreement Drift in LLM Debates 研究LLM在不平衡辩论中的行为，揭示网络效应和“一致性漂移”现象 large language model
26 PaperScope: A Multi-Modal Multi-Document Benchmark for Agentic Deep Research Across Massive Scientific Papers 提出PaperScope：一个用于评估Agentic深度研究的多模态多文档基准。 large language model
27 ZoomR: Memory Efficient Reasoning through Multi-Granularity Key Value Retrieval ZoomR：通过多粒度键值检索实现内存高效的LLM推理 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (12 篇)

#题目一句话要点标签🔗⭐
28 A Mamba-Based Multimodal Network for Multiscale Blast-Induced Rapid Structural Damage Assessment 提出基于Mamba的多模态网络，用于爆炸诱导的多尺度快速结构损伤评估 Mamba multimodal ✅
29 From Topology to Trajectory: LLM-Driven World Models For Supply Chain Resilience 提出 ReflectiChain，利用LLM驱动的世界模型提升供应链韧性 world model world models large language model
30 Escaping the Context Bottleneck: Active Context Curation for LLM Agents via Reinforcement Learning 提出基于强化学习的主动上下文管理框架，解决LLM Agent长程任务中的上下文瓶颈问题 reinforcement learning large language model foundation model
31 Collaborative Multi-Agent Scripts Generation for Enhancing Imperfect-Information Reasoning in Murder Mystery Games 提出协同多智能体剧本生成框架，提升VLMs在谋杀之谜游戏中不完美信息推理能力 reinforcement learning reward shaping multimodal
32 CSPO: Alleviating Reward Ambiguity for Structured Table-to-LaTeX Generation 提出CSPO框架，缓解结构化表格转LaTeX生成中的奖励模糊问题。 reinforcement learning large language model multimodal
33 CoRe-ECG: Advancing Self-Supervised Representation Learning for 12-Lead ECG via Contrastive and Reconstructive Synergy 提出CoRe-ECG以解决ECG自监督学习中的数据稀缺问题 representation learning contrastive learning
34 OOM-RL: Out-of-Money Reinforcement Learning Market-Driven Alignment for LLM-Based Multi-Agent Systems 提出OOM-RL，利用金融市场损耗作为负梯度，对LLM多智能体系统进行市场驱动的对齐。 reinforcement learning RLHF
35 MADQRL: Distributed Quantum Reinforcement Learning Framework for Multi-Agent Environments 提出MADQRL：一种用于多智能体环境的分布式量子强化学习框架 reinforcement learning
36 Three Roles, One Model: Role Orchestration at Inference Time to Close the Performance Gap Between Small and Large Agents 提出基于角色编排的推理时框架，提升小模型Agent在复杂任务中的性能 reinforcement learning large language model
37 MAFIG: Multi-agent Driven Formal Instruction Generation Framework 提出MAFIG框架以解决调度系统应急处理问题 distillation large language model
38 Deep Learning for Sequential Decision Making under Uncertainty: Foundations, Frameworks, and Frontiers 综述：深度学习赋能不确定性下的序贯决策，弥合AI与运筹优化 reinforcement learning deep reinforcement learning
39 NimbusGuard: A Novel Framework for Proactive Kubernetes Autoscaling Using Deep Q-Networks NimbusGuard：利用深度Q网络实现Kubernetes主动式自动伸缩 reinforcement learning deep reinforcement learning

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

#题目一句话要点标签🔗⭐
40 Taking a Pulse on How Generative AI is Reshaping the Software Engineering Research Landscape 大规模调研揭示生成式AI对软件工程研究的重塑：使用、影响与治理 PULSE

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
41 AI Integrity: A New Paradigm for Verifiable AI Governance 提出AI完整性以解决AI治理中的可验证性问题 manipulation

⬅️ 返回 cs.AI 首页 · 🏠 返回主页