cs.AI（2026-01-06）

📊 共 20 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) (16 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (4)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (16 篇)

#题目一句话要点标签🔗⭐
1 Rationale-Grounded In-Context Learning for Time Series Reasoning with Multimodal Large Language Models 提出RationaleTS，通过先验知识推理提升多模态大语言模型在时间序列推理中的表现 large language model multimodal
2 Multi-Modal Data-Enhanced Foundation Models for Prediction and Control in Wireless Networks: A Survey 综述：多模态数据增强的无线网络预测与控制基础模型 foundation model
3 MoE Adapter for Large Audio Language Models: Sparsity, Disentanglement, and Gradient-Conflict-Free 提出MoE-Adapter，解决大语音语言模型中音频异构性导致的梯度冲突问题。 large language model multimodal
4 LLM Agent Framework for Intelligent Change Analysis in Urban Environment using Remote Sensing Imagery 提出ChangeGPT：基于LLM Agent的遥感影像城市环境智能变化分析框架 large language model foundation model
5 Learning from Prompt itself: the Hierarchical Attribution Prompt Optimization 提出HAPO框架，解决提示词优化中的漂移和可解释性问题，提升多模态任务性能。 large language model multimodal
6 MAGMA: A Multi-Graph based Agentic Memory Architecture for AI Agents 提出MAGMA：一种基于多图的Agentic记忆架构，用于提升AI Agent的长程推理能力。 large language model
7 Fine-tuning Small Language Models as Efficient Enterprise Search Relevance Labelers 提出一种高效微调小语言模型的企业搜索相关性标注方法，媲美甚至超越大模型。 large language model
8 JPU: Bridging Jailbreak Defense and Unlearning via On-Policy Path Rectification 提出JPU，通过在线策略路径修正桥接越狱防御与模型卸载 large language model
9 Batch-of-Thought: Cross-Instance Learning for Enhanced LLM Reasoning 提出Batch-of-Thought，通过跨实例学习增强LLM推理能力 large language model
10 Logical Phase Transitions: Understanding Collapse in LLM Logical Reasoning 揭示LLM逻辑推理的相变现象，提出神经符号课程调优方法 large language model ✅
11 ReTreVal: Reasoning Tree with Validation -- A Hybrid Framework for Enhanced LLM Multi-Step Reasoning ReTreVal：融合验证的推理树，增强LLM多步推理能力 large language model
12 M3MAD-Bench: Are Multi-Agent Debates Really Effective Across Domains and Modalities? M3MAD-Bench：多智能体辩论在跨领域和跨模态场景下的有效性评估基准 multimodal ✅
13 Netflix Artwork Personalization via LLM Post-training 通过LLM后训练实现Netflix艺术作品个性化推荐，提升用户满意度。 large language model
14 The Path Ahead for Agentic AI: Challenges and Opportunities 探索Agentic AI的未来：挑战与机遇，聚焦架构演进与技术瓶颈 large language model
15 Hypothesize-Then-Verify: Speculative Root Cause Analysis for Microservices with Pathwise Parallelism SpecRCA：基于假设-验证范式的微服务推测性根因分析框架 large language model
16 TAAF: A Trace Abstraction and Analysis Framework Synergizing Knowledge Graphs and LLMs TAAF：结合知识图谱与LLM的追踪抽象与分析框架，提升系统理解 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (4 篇)

#题目一句话要点标签🔗⭐
17 Interpretable All-Type Audio Deepfake Detection with Audio LLMs via Frequency-Time Reinforcement Learning 提出基于频率-时间强化学习的音频大语言模型，用于可解释的全类型音频深度伪造检测 reinforcement learning large language model chain-of-thought
18 Sample-Efficient Neurosymbolic Deep Reinforcement Learning 提出神经符号深度强化学习方法，提升样本效率和泛化能力，解决复杂环境下的序贯决策问题。 reinforcement learning deep reinforcement learning DRL
19 Time-Scaling Is What Agents Need Now 提出时间尺度调整，提升智能体在认知约束下的深度推理与问题解决能力 reinforcement learning world model large language model
20 Inferring Causal Graph Temporal Logic Formulas to Expedite Reinforcement Learning in Temporally Extended Tasks 提出GTL-CIRL框架，通过学习因果图时序逻辑公式加速时序扩展任务中的强化学习。 reinforcement learning

⬅️ 返回 cs.AI 首页 · 🏠 返回主页