cs.AI（2026-01-27）

📊 共 29 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) (21 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (6 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (1) 支柱七：动作重定向 (Motion Retargeting) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (21 篇)

#题目一句话要点标签🔗⭐
1 A Benchmark for Audio Reasoning Capabilities of Multimodal Large Language Models 提出音频推理任务基准(ART)，用于评估多模态大语言模型的音频推理能力 large language model multimodal
2 TS-Debate: Multimodal Collaborative Debate for Zero-Shot Time Series Reasoning 提出TS-Debate，用于零样本时间序列推理的多模态协同辩论框架 large language model multimodal
3 SAM Audio Judge: A Unified Multimodal Framework for Perceptual Evaluation of Audio Separation 提出SAM Audio Judge，用于无需参考信号的多模态音频分离质量评估 multimodal ✅
4 CoReTab: Improving Multimodal Table Understanding with Code-driven Reasoning 提出CoReTab框架，通过代码驱动推理提升多模态表格理解能力。 multimodal
5 GAVEL: Towards rule-based safety through activation monitoring GAVEL：提出基于激活监控的规则化安全框架，提升LLM安全性。 large language model
6 Benchmarks Saturate When The Model Gets Smarter Than The Judge Omni-MATH-2：通过高质量数据集和可靠评估，提升数学问题基准测试的准确性。 large language model
7 RPO:Reinforcement Fine-Tuning with Partial Reasoning Optimization RPO：通过部分推理优化进行强化微调，显著降低计算开销。 large language model ✅
8 Beyond In-Domain Detection: SpikeScore for Cross-Domain Hallucination Detection 提出SpikeScore，用于解决大语言模型跨领域幻觉检测问题 large language model
9 AgenticSCR: An Autonomous Agentic Secure Code Review for Immature Vulnerabilities Detection AgenticSCR：一种用于检测早期漏洞的自主智能体安全代码审查方法 large language model
10 Veri-Sure: A Contract-Aware Multi-Agent Framework with Temporal Tracing and Formal Verification for Correct RTL Code Generation 提出Veri-Sure框架，通过形式化验证提升LLM在RTL代码生成中的正确性。 large language model
11 RvB: Automating AI System Hardening via Iterative Red-Blue Games RvB：通过迭代红蓝对抗博弈自动化AI系统强化 large language model
12 Algorithmic Prompt-Augmentation for Efficient LLM-Based Heuristic Design for A* Search 提出算法提示增强的A-CEoH框架，高效利用LLM为A*搜索设计启发式函数。 large language model
13 ComAgent: Multi-LLM based Agentic AI Empowered Intelligent Wireless Networks ComAgent：基于多LLM智能体的智能无线网络方案，提升跨层优化效率 large language model
14 AACR-Bench: Evaluating Automatic Code Review with Holistic Repository-Level Context 提出AACR-Bench，用于在代码仓库级别评估自动化代码审查中的大型语言模型 large language model ✅
15 Revisiting Parameter Server in LLM Post-Training 针对LLM后训练中负载不均衡问题，提出On-Demand Communication加速训练。 large language model ✅
16 Balancing Sustainability And Performance: The Role Of Small-Scale Llms In Agentic Artificial Intelligence Systems 利用小规模LLM平衡Agentic AI系统的可持续性与性能 large language model
17 GLOVE: Global Verifier for LLM Memory-Environment Realignment GLOVE：用于LLM记忆-环境重对齐的全局验证器 large language model
18 MAGNET: Towards Adaptive GUI Agents with Memory-Driven Knowledge Evolution 提出MAGNET，利用记忆驱动知识演化实现自适应GUI代理 foundation model
19 Multi-Agent Procedural Graph Extraction with Structural and Logical Refinement 提出多代理程序图提取框架以解决结构有效性与逻辑一致性问题 large language model
20 Detecting and Correcting Hallucinations in LLM-Generated Code via Deterministic AST Analysis 提出基于确定性AST分析的代码幻觉检测与纠正框架，提升LLM代码生成可靠性 large language model
21 HalluJudge: A Reference-Free Hallucination Detection for Context Misalignment in Code Review Automation 提出HalluJudge，用于检测代码评审自动化中上下文不一致导致的幻觉问题 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🔗⭐
22 Visual Generation Unlocks Human-Like Reasoning through Multimodal World Models 提出视觉优越性假设，利用多模态世界模型提升类人推理能力 world model large language model multimodal
23 Exploring Weaknesses in Function Call Models via Reinforcement Learning: An Adversarial Data Augmentation Approach 提出基于强化学习的对抗数据增强方法，提升函数调用大语言模型的鲁棒性 reinforcement learning large language model
24 Residual Tokens Enhance Masked Autoencoders for Speech Modeling RT-MAE：利用残差令牌增强语音建模的掩码自编码器 masked autoencoder MAE
25 Agentic Design Patterns: A System-Theoretic Framework 提出基于系统论的Agentic设计模式框架，解决智能体系统设计中的不可靠和脆弱性问题。 world model foundation model
26 AlignCoder: Aligning Retrieval with Target Intent for Repository-Level Code Completion AlignCoder：通过对齐检索与目标意图实现仓库级代码补全 reinforcement learning large language model
27 Uncertainty-Aware 3D Emotional Talking Face Synthesis with Emotion Prior Distillation 提出UA-3DTalk，解决3D情感语音人脸合成中的情感对齐和渲染质量问题。 distillation ✅

🔬 支柱三：空间感知与语义 (Perception & Semantics) (1 篇)

#题目一句话要点标签🔗⭐
28 Bridging Visual and Wireless Sensing: A Unified Radiation Field for 3D Radio Map Construction 提出URF-GS框架，融合视觉与无线感知，构建精确可泛化的3D无线电地图 3D gaussian splatting gaussian splatting splatting

🔬 支柱七：动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗⭐
29 M-SGWR: Multiscale Similarity and Geographically Weighted Regression 提出M-SGWR以解决传统地理加权回归模型的局限性 spatial relationship

⬅️ 返回 cs.AI 首页 · 🏠 返回主页