cs.AI（2026-10-29）

📊 共 45 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) (29 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (14 🔗1) 支柱一：机器人控制 (Robot Control) (1 🔗1) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (29 篇)

#题目一句话要点标签🔗⭐
1 Distributional Reward Calibration and Budget-Neutral Credit Assignment for Multimodal RLVR 提出TACE以解决多模态强化学习中的奖励校准问题 multimodal
2 Category-Anchored Large Language Models for Next POI Recommendation 提出ANCHOR框架以解决POI推荐中的语义与实体间隔问题 large language model
3 SocialArena: Evaluating Social Intelligence in Large Language Models 提出SocialArena以系统评估大型语言模型的社交智能 large language model
4 Only Accepted Tokens Count: An Energy Accounting of Speculative Decoding Under Latency Targets 提出能量核算方法以优化推理延迟目标下的语言模型解码 large language model
5 Understanding axial attention in TSFMs through single location regression 扩展单位置回归框架以研究轴向注意力机制 foundation model
6 TextSeal2: Distortion-Free LLM Watermarking with Copula Mixing and Memoryless Races 提出TextSeal2以解决大语言模型水印失真问题 large language model
7 Comparing Human Role-Players and LLM-Simulated Clients in Online Counselling Training: An Analysis of Counselling Patterns 比较人类角色扮演者与LLM模拟客户在在线咨询培训中的表现 large language model
8 Agentic University: Collaborative Agents for Personalized Learning 提出Agentic University以解决个性化学习中的协作与评估问题 large language model ✅
9 UPPAAi: Towards LLM-Assisted Modeling, Verification, Analysis and Synthesis 提出UPPAAI以解决实时与网络物理系统建模难题 large language model
10 HEMO: Harness Self-Evolution for Multi-Objective Combinatorial Optimization 提出HEMO以解决多目标组合优化中的自我演化问题 large language model
11 InferD: A Honeypot Framework for AI Inference Infrastructure 提出InferD框架以解决AI推理基础设施的蜜罐监测问题 large language model
12 Finding Is Not Returning, Returning Is Not Reconstructing: A Multi-Dimensional Evaluation of Agent Social Re-entry 提出多维评估框架以解决智能体社交重返问题 large language model
13 The Real Cost of Context: Predicting Compute Time and Energy Consumption in LLM Prefill and Decode Phases 提出计算时间和能耗预测模型以优化LLM的上下文处理 large language model
14 Non-Zero Radius: The Missing Component of Neural Networks 提出球形神经网络以解决符号推理能力不足问题 large language model
15 Verifiable LLM-Guided Focal SMT Solving for Quantified Arrays 提出LinguaArray以解决量化数组SMT求解问题 large language model
16 Well-Specified and Well-Tested? Assessing SWE-bench-Style Datasets with a Self-Evolving Agent 提出ExLabel以解决SWE数据集质量评估问题 large language model
17 OPERA: OPamp Design by an Expert-Taught Reasoning Agent, from Specifications to Layout 提出OPERA以实现从规格到布局的CMOS运算放大器设计 large language model ✅
18 What Does "Now" Mean to Your Computer-Use Agent? Stale Screenshots, Page Loads, and Which Check to Run Before a Click 提出基于像素比较的检查机制以解决点击失效问题 multimodal ✅
19 LLM-Assisted Optimization for Simulation-Based Redesign 提出LLM辅助优化方法以解决仿真重设计问题 large language model
20 HERA: Assessing an Autonomous Agent for Quantum Device Measurement 提出HERA以解决量子设备测量的自主代理问题 large language model
21 Incongruence-Aware Evaluation of Multimodal Speech Emotion Recognition in Audio Large Language Models 提出一种不一致性评分以提升多模态语音情感识别准确性 large language model multimodal
22 Opening the Box Changes the Cat: LLMs Detect Chain-of-Thought Monitoring - What now? 探讨LLMs在链式思维监控中的适应性行为 large language model chain-of-thought
23 The psychophysics of affect in large language models 探讨大语言模型中情感映射与决策的关系 large language model
24 The First Few Tokens Matter: Understanding Chain-of-Thought Controllability 提出基于初始令牌干预的链式思维可控性提升方法 chain-of-thought
25 Accept Changes, Reject Attacks: Robust Memory Management for LLM Agents 提出一种鲁棒内存管理方法以解决LLM代理的记忆问题 large language model
26 P-ContextBench: From Personal Fact Recall to Personalized Agent Behavior 提出P-ContextBench以评估个人上下文理解能力 large language model
27 Towards Portable Specification of the Agent Harness 提出可移植的代理工具规范以解决模型评估不一致问题 large language model
28 Safin-1.5: Safety-Native Architecture and Verifiable Reasoning for Recursive Safety Improvement 提出Safin-1.5以解决递归自我改进中的安全性问题 large language model
29 CLIC: Closed-Loop Intent Consistency for Securing LLM-Driven Control in Open RAN 提出闭环意图一致性层以解决O-RAN控制中的安全问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (14 篇)

#题目一句话要点标签🔗⭐
30 Revisiting Over-refusal in Safety-aligned Large Language Models: A Representation and Optimization Perspective 提出GOAL框架以解决大型语言模型中的过度拒绝问题 contrastive learning large language model
31 Timbre: Reliability-aware harmonic consensus for label-free multimodal test-time rexinforcement learning 提出可靠性感知的谐波共识框架解决无标签多模态强化学习问题 reinforcement learning multimodal
32 Algebraic World Model: From Exploration to Proofs 提出代数世界模型以解决未知环境中的推理问题 world model world models
33 Proving and Improving the Platonic Representation Hypothesis with JEPAs 提出JEPAs以证明和改进柏拉图表征假设 JEPA Joint-Embedding Predictive Architecture joint-embedding predictive architecture
34 Online Dense Reinforcement Learning with Planning-Grounding Agent for UI Navigation on Live IOS Apps 提出在线密集强化学习框架以解决iOS应用UI导航问题 reinforcement learning PPO offline reinforcement learning
35 Cheap Retrieval and Observation Credit Assignment from One Quantity: The Value of Information 提出基于信息价值的记忆检索与观察信用分配方法 reinforcement learning world model world models
36 Sandman: Building and Operating Sandboxes for Agentic Reinforcement Learning on Shared Kubernetes 提出Sandman以解决共享Kubernetes环境下的强化学习沙箱管理问题 reinforcement learning
37 On the Fundamental Limits of Prefill-Based Decoding Length Prediction 提出基于预填充阶段的解码长度预测方法以解决推理调度问题 MAE large language model ✅
38 EnvOS: Auditing Computer-Use Agents When the World Moves On : Phenomenon-Based Task Construction, Certification, and Evidence-Based Evaluation 提出EnvOS框架以解决动态环境下计算机代理的审计问题 reinforcement learning large language model
39 Integrated Feature Selection for classification and Representation Learning with Group-Lasso Siamese Networks 提出基于Group-Lasso的Siamese网络进行特征选择与表示学习 representation learning
40 Reasoning or Pattern Exploitation? Mechanistic Insights from RL-Trained Language Models 提出机制性洞察以区分RL训练语言模型的推理与模式利用 reinforcement learning large language model
41 Agentic World Modeling for Agent Test-time Scaling without Environment Replica 提出Agentic世界建模以解决共享环境下的测试时间扩展问题 world model world models large language model
42 Beyond Plausible Prediction: Safety Sufficiency in Web World Models 提出安全充分性评估框架以解决网络世界模型的安全决策问题 world model world models
43 Algebraic World Model: From Exploration to Proofs 提出代数世界模型以解决未知环境中的推理问题 world model world models

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
44 AegisFlow: Training-Free Non-Myopic Path-Safe Guided Flow Matching 提出AegisFlow以解决生成模型路径安全性问题 manipulation flow matching ✅

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

#题目一句话要点标签🔗⭐
45 Collaborative AI agents for autonomous experimental research with solid-state quantum spins 提出协作AI代理以实现自主实验研究 PULSE large language model

⬅️ 返回 cs.AI 首页 · 🏠 返回主页