cs.AI（2026-01-20）

📊 共 25 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) (17) 支柱二：RL算法与架构 (RL & Architecture) (5 🔗2) 支柱八：物理动画 (Physics-based Animation) (2 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (17 篇)

#	题目	一句话要点	标签	🔗	⭐
1	SilentDrift: Exploiting Action Chunking for Stealthy Backdoor Attacks on Vision-Language-Action Models	SilentDrift：利用动作分块对VLA模型进行隐蔽后门攻击	vision-language-action VLA
2	Diffusion Large Language Models for Black-Box Optimization	提出基于扩散语言模型的黑盒优化方法dLLM，在少量样本下实现设计优化。	large language model
3	Measuring the State of Open Science in Transportation Using Large Language Models	利用大型语言模型评估交通运输研究中的开放科学实践现状	large language model
4	On the Generalization Gap in LLM Planning: Tests and Verifier-Reward RL	揭示LLM规划泛化差距：提出诊断干预方法与验证器奖励强化学习	large language model
5	VisTIRA: Closing the Image-Text Modality Gap in Visual Math Reasoning via Structured Tool Integration	VisTIRA：通过结构化工具集成弥合视觉数学推理中的图像-文本模态差距	chain-of-thought
6	Opportunities in AI/ML for the Rubin LSST Dark Energy Science Collaboration	探索AI/ML在Rubin LSST暗能量科学合作中的应用机遇与挑战	foundation model
7	Human Simulation Computation: A Human-Inspired Framework for Adaptive AI Systems	提出人类模拟计算框架HSC，提升AI系统在动态环境中的适应性和推理能力	large language model
8	LifeAgentBench: A Multi-dimensional Benchmark and Agent for Personal Health Assistants in Digital Health	提出LifeAgentBench，用于评估和提升数字健康中个人健康助手的能力。	large language model
9	HardSecBench: Benchmarking the Security Awareness of LLMs for Hardware Code Generation	HardSecBench：评估LLM在硬件代码生成中的安全意识基准	large language model
10	ToolCaching: Towards Efficient Caching for LLM Tool-calling	提出ToolCaching，解决LLM工具调用中冗余请求问题，提升缓存效率。	large language model
11	Hidden in Plain Text: Measuring LLM Deception Quality Against Human Baselines Using Social Deduction Games	利用社交推理游戏评估LLM在自然语言欺骗中的表现，优于人类基线	large language model
12	Why Does the LLM Stop Computing: An Empirical Study of User-Reported Failures in Open-Source LLMs	大规模实证研究开源LLM部署失败问题，揭示系统性瓶颈与解决方案。	large language model
13	Foundations of Global Consistency Checking with Noisy LLM Oracles	提出基于LLM的全局一致性检查框架，通过自适应分治算法高效检测并修复不一致性。	large language model
14	DSAEval: Evaluating Data Science Agents on a Wide Range of Real-World Data Science Problems	DSAEval：提出一个真实世界数据科学问题评估基准，用于评估数据科学Agent的性能。	multimodal
15	SCRIPTMIND: Crime Script Inference and Cognitive Evaluation for LLM-based Social Engineering Scam Detection System	ScriptMind：用于LLM社交工程诈骗检测的犯罪脚本推理与认知评估框架	large language model
16	Leveraging ChatGPT and Other NLP Methods for Identifying Risk and Protective Behaviors in MSM: Social Media and Dating apps Text Analysis	利用ChatGPT等NLP方法识别MSM人群的风险和保护行为：社交媒体和约会应用文本分析	large language model
17	CatMaster: An Agentic Autonomous System for Computational Heterogeneous Catalysis Research	CatMaster：基于LLM的自主智能体系统，加速计算异构催化研究	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	🔗	⭐
18	Large Language Model-Powered Evolutionary Code Optimization on a Phylogenetic Tree	提出PhyloEvolve以解决GPU算法优化中的效率问题	reinforcement learning decision transformer distillation	✅
19	"Just in Time" World Modeling Supports Human Planning and Reasoning	提出“即时”世界建模框架，支持人类规划与推理	world model
20	Toward Efficient Agents: Memory, Tool learning, and Planning	针对Agent系统效率瓶颈，提出内存优化、工具学习和规划的综合改进方案	reinforcement learning large language model
21	DARC: Decoupled Asymmetric Reasoning Curriculum for LLM Evolution	DARC：解耦非对称推理课程，提升大型语言模型自进化能力	distillation large language model	✅
22	Learning Discrete Successor Transitions in Continuous Attractor Networks: Emergence, Limits, and Topological Constraints	研究连续吸引子网络中离散后继状态转移的学习：涌现、局限与拓扑约束	curriculum learning PULSE

🔬 支柱八：物理动画 (Physics-based Animation) (2 篇)

#	题目	一句话要点	标签	🔗	⭐
23	Resilient Routing: Risk-Aware Dynamic Routing in Smart Logistics via Spatiotemporal Graph Learning	提出基于时空图学习的风险感知动态路径规划框架，提升智能物流韧性。	spatiotemporal
24	torch-sla: Differentiable Sparse Linear Algebra with Adjoint Solvers and Sparse Tensor Parallelism for PyTorch	提出torch-sla以解决稀疏线性代数计算效率问题	differentiable simulation	✅

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

#	题目	一句话要点	标签	🔗	⭐
25	MATE: Matryoshka Audio-Text Embeddings for Open-Vocabulary Keyword Spotting	提出MATE：用于开放词汇关键词检索的俄罗斯套娃式音频-文本嵌入	open-vocabulary open vocabulary

⬅️ 返回 cs.AI 首页 · 🏠 返回主页