cs.AI（2026-09-28）

📊 共 38 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 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (11 🔗2)

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

#题目一句话要点标签🔗⭐
1 Textual User Taste: Natural-Language User Context for Foundation-Model Recommender System at Scale 提出文本用户品味以优化推荐系统用户上下文问题 large language model foundation model
2 Signatures of semantic search in the activations of large language models 提出机制可解释性技术以揭示大语言模型的语义搜索特征 large language model
3 Tool Mediation Alters Refusal Mechanisms in Large Language Models 探讨工具介导如何改变大型语言模型的拒绝机制 large language model
4 What Drives Citations in Production Large Language Models? An Observational Multi-Method Study of Two Million AI Citations Across Ten Thousand Web Pages 提出多方法研究以揭示生产大型语言模型引用驱动因素 large language model
5 SeLMRoute: Probabilistic Semantic Evidence for Large Language Model Routing 提出SeLMRoute以优化大语言模型路由问题 large language model ✅
6 From One-Shot Generation to Incremental Music Composition: Adapting a General-Purpose Instruction LLM for Persistent Symbolic Editing 提出基于LLM的增量音乐创作方法以解决传统生成系统局限 large language model instruction following
7 Not All Thinking is Created Equal: Latent Reasoning Discovers a Recurrent Search Algorithm for Depth Generalization 提出潜在推理机制以解决深度泛化问题 large language model chain-of-thought
8 AUV-Bench: Aesthetic Understanding and Generation Evaluation for User Interfaces 提出AUV-Bench以解决用户界面美学评估与生成问题 foundation model multimodal
9 QC-Stark: A Multi-Task Benchmark Revealing Capability Dissociations in LLMs Evaluated on Quantum Computing Tasks 提出QC-Stark基准以评估LLMs在量子计算任务中的能力差异 large language model
10 Share-Borne AI Virus: Memory-Hopping Attacks Across LLM Agents 提出共享AI病毒以解决大语言模型的安全隐患问题 large language model
11 SRHarness: A Harness for Agentic Symbolic Regression 提出SRHarness以提升代理符号回归的性能 large language model
12 TempoKV: Timely Staging of LLM KV Caches for Memory-Semantic Flash 提出TempoKV以优化LLM KV缓存的内存管理问题 large language model
13 TCSAlgBench: Benchmarking Automated Proving for Research-Level Theoretical Computer Science 提出TCSAlgBench以评估自动证明在理论计算机科学中的应用 large language model
14 SEABench: Benchmarking Endogenous Misalignment In Self-Evolving Agents 提出SEABench以研究自演化智能体中的内生性不对齐问题 chain-of-thought
15 IMC-CLINIC: Coupled Loss-Informed Newton Iterations for Clipping in Analog In-Memory Computing 提出IMC-CLINIC以解决模拟内存计算中的剪切优化问题 large language model
16 "Nothing to See Here'': Unintended Disclosure through Revision Traces of LLM Deliverables 提出RevLeakBench以解决LLM交互中的意外信息泄露问题 large language model
17 Jev thinks "I don't know'', but doesn't say it: Introducing Sys1Cal-v1 Dataset for Probability Calibration 提出Sys1Cal-v1数据集以解决概率校准问题 foundation model
18 Narrowing the Horizon: Quantifying Topic Saliency Shifts in Generative Monoculture 提出主题显著性测量方法以应对生成单一文化问题 large language model ✅
19 Towards Reliable AI Data Scientists: Data Agents with Workflow Harnesses 提出数据代理与工作流框架以解决数据科学可靠性问题 large language model ✅
20 EP-Mem: Elastic Privacy Memory for Social Relationship-Aware LLM Agents 提出EP-Mem以解决社交关系感知LLM代理的隐私问题 large language model
21 Beneath the Tokens: A Performance Engineering Study of Multi-Token Prediction in GPU-Accelerated LLM Inference 提出多令牌预测以提升GPU加速LLM推理性能 large language model
22 Persona Following Is Not Selective Control: The Neutrality Gap in LLM User Simulation 提出中立性差距以解决大语言模型用户模拟中的选择性控制问题 large language model
23 Automated feature engineering, AutoML, and decision-focused learning for improved energy consumption forecasting 提出自动特征工程以提升能源消费预测精度 TAMP
24 Action-Space Shaping for LLM Agents: Measuring and Mitigating Tool-Schema Bias 提出可执行变换框架以解决工具模式偏差问题 large language model
25 PDEU-Bench: Benchmarking the Personalized Planning Lifecycle of Tool-Calling LLM Agents 提出PDEU-Bench以评估个性化工具调用LLM代理的规划生命周期 large language model
26 One Readout, Many Repairs: Diffusion-Guided Hierarchical Search for Tool-Agent Repair 提出ReCommit框架以解决工具代理修复问题 large language model
27 PainterBench: A Figural Divergent-Thinking Benchmark for Tool-Using Language Models 提出PainterBench以评估工具使用语言模型的图形发散思维能力 multimodal

🔬 支柱二：RL算法与架构 (RL & Architecture) (11 篇)

#题目一句话要点标签🔗⭐
28 Privacy-Preserving Full-Body Meshing from mmWave Radar via Mesh Foundation Model Supervision 提出跨模态教师-学生框架以实现隐私保护的全身3D网格重建 teacher-student sam 3D SAM 3D
29 A.D.A.M.O. (Agent for language-Driven Actions with Multimodal Observations): A Visual-Symbolic Framework for Virtual Humans 提出A.D.A.M.O.框架以解决虚拟人语言驱动行为的整合问题 world model world models egocentric
30 ARISE: Adapting to Evolving Capability Gaps in Agentic Reinforcement Learning 提出ARISE框架以解决智能体强化学习中的能力差距问题 reinforcement learning foundation model ✅
31 ASCT: Attentive Search over Counterfactual Trees for Credit Assignment in Agentic Reinforcement Learning 提出ASCT框架以解决代理强化学习中的信用分配问题 reinforcement learning policy learning PPO
32 Can We Trust the Teacher? Decoupled Credit Direction-Magnitude for Self-Distillation 提出解耦信用方向与幅度的自蒸馏方法以提升策略优化 distillation multimodal
33 Continuous Context Management 提出连续上下文管理方法以优化长时交互历史处理 reinforcement learning distillation large language model
34 Using Context Is Not Enough: Test-Time Training for Personalized Reward Modeling 提出P-TTT以解决个性化奖励建模中的偏好关系捕捉问题 reinforcement learning RLHF large language model
35 Safe Greenhouse Climate Control Using Lagrangian-Constrained PPO with Kolmogorov-Arnold Networks 提出基于Lagrangian约束的PPO与Kolmogorov-Arnold网络的温室气候控制方法 reinforcement learning PPO
36 Reinforcing Agentic Creativity in Scientific Ideation with Night Science 提出AI Night-Scientist以增强科学创意的多样性 reinforcement learning large language model
37 PEARL: Adaptive Prefill-Decode Execution with Elasticity for Agentic Reinforcement Learning 提出PEARL以优化智能强化学习中的资源利用与执行效率 reinforcement learning
38 JRDB-AVR: An Active Visual Reasoning Benchmark for Embodied Agents in Real-World Environments 提出JRDB-AVR以解决复杂环境中主动视觉推理问题 world model world models ✅

⬅️ 返回 cs.AI 首页 · 🏠 返回主页