cs.AI（2026-09-24）

📊 共 40 篇论文 | 🔗 7 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (25 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (13 🔗3) 支柱一：机器人控制 (Robot Control) (2 🔗2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (25 篇)

#题目一句话要点标签🔗⭐
1 C3M: Cross-Session Multimodal Memory Maintenance for Long-Horizon Tasks 提出C3M以解决长时间任务中的跨会话多模态记忆维护问题 multimodal ✅
2 Style, Not Self: Surface Cues Explain Zero-Shot Code Attribution by Large Language Models 提出基于表面线索的零-shot代码归属方法以解决模型自我偏好问题 large language model
3 SkinAgent AI: A Safety-Grounded Multimodal Agentic Framework for Non-Diagnostic Skincare Support 提出SkinAgent AI框架以解决非诊断性护肤支持问题 multimodal
4 When Can Agents Forget Their Reasoning? ICLR for Long-Horizon Agent Context Compression 提出互动感知压缩方法以解决长时间推理中的历史信息遗忘问题 chain-of-thought
5 GRASP: Generating, Revising, and Assessing for Strategic Planning with Agentic AI 提出GRASP框架以解决复杂任务规划问题 large language model
6 A Rapid Pipeline for Training and Deploying ML Models on WeBe Band 提出一种快速管道以优化和部署边缘设备上的机器学习模型 multimodal
7 SAGE: Mitigating Long-Horizon Reasoning Biases via Topological Guidance 提出SAGE以解决长远推理中的偏差问题 large language model ✅
8 Screen Before You Serve: Simulation for Production Customer Experience AI Agents at 140M Scale 提出一种模拟工作流程以优化客户体验AI代理的部署 large language model
9 PrivDrift: Auditing User-Secret Leakage Under Topic Drift in Active LLM Conversations 提出PrivDrift以审计用户秘密在对话主题漂移下的泄露问题 large language model
10 Guardrails or Roadblocks? Effects of Pedagogical Style and Context Awareness in AI Teaching Assistants for Programming 探讨AI教学助手的教学风格与上下文意识对编程学习的影响 large language model
11 Neuro-symbolic AI for Industrial Configuration 提出神经符号AI以解决工业产品配置问题 large language model
12 Who Holds the Pen? Let Specifications, Not Agents, Sign Off 提出规范主导的执行框架以解决智能体执行中的理解与权威差距问题 large language model
13 Hallucination Neurons and Where to Find Them: An Investigation into the existence of Hallucination Neurons 提出五步诊断协议以解决稀疏神经元定位问题 large language model
14 Prefilling the Reasoning Channel: Output-Prefix Attacks on Reasoning LLMs 提出输出前缀攻击以破解推理型大型语言模型 large language model
15 SWE-Prometheus: Measuring Engineering Governance Improvements in Real-World Repositories 提出SWE-Prometheus以提升软件工程治理评估 large language model
16 Epistemic-Probabilistic Model for Guarded Multi-Agent LLM Coordination 提出一种新型模型以解决多智能体LLM协调问题 large language model
17 Spot, Separate, and Enhance: Fully Generative Approach for Audio Mixing 提出SSE以解决音频混音与增强问题 multimodal
18 A Wrong Turn Does Not Ruin the Journey: Deviation-Guided Skill Self-Evolution for LLM Agents 提出SkillPivot框架以解决LLM代理技能自我演化问题 large language model
19 From Static Personal Values to Contextualized Personalization: Bayesian Personalized Value Alignment for LLMs 提出BaCVA以解决大语言模型个性化价值对齐问题 large language model
20 Human-AI-Powered Hypothesis Testing: Cost-Aware Selective AI Scoring and Sequential Human Escalation 提出SCALE以降低人机假设检验成本 large language model
21 BioEVAL: A global, multi-institutional benchmark of large language and multimodal models for bioengineering 提出BioEVAL基准以评估生物工程领域的语言与多模态模型 large language model multimodal
22 T-RoPE: Time-Aware Rotary Position Embedding for Sequential Recommendation 提出T-RoPE以解决推荐系统中的时间感知问题 large language model TAMP
23 Audio LLMs Know When They Can't Hear You 提出音频LLM自我识别转录可靠性的方法以解决误解问题 large language model
24 PrivDrift: Auditing User-Secret Leakage Under Topic Drift in Active LLM Conversations 提出PrivDrift以审计主动LLM对用户秘密泄露的风险 large language model
25 Bridging LLM Agents and Data Spaces: An Architectural Mediation Approach using the Model Context Protocol 提出模型上下文协议以解决LLM代理与数据空间集成问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (13 篇)

#题目一句话要点标签🔗⭐
26 Calibrated Decision Models for Autonomous Penetration-Testing Harnesses: JEV and Laya as System One Decision Layers for LLM-Driven Pentest Agents 提出系统一决策模型以提升自主渗透测试的准确性 RLHF penetration large language model
27 Search-Aware Reinforcement Learning for Multi-Component Query Understanding in Roblox Game Search 提出搜索感知强化学习以优化Roblox游戏搜索的查询理解 reinforcement learning teacher-student large language model
28 From Self-Distillation to Self-Practice: Privileged Information for Multi-Turn Agents 提出特权自实践方法以解决多回合智能体的表现不足问题 distillation privileged information
29 RD-JEPA: Predictive latent pretraining for few-trajectory transfer across reaction--diffusion equations 提出RD-JEPA以解决反应-扩散方程的少量轨迹迁移问题 JEPA Joint-Embedding Predictive Architecture joint-embedding predictive architecture
30 Back to the Definition: Estimating Step-Level Advantages via Trajectory Graphs for Agentic Reinforcement Learning 提出GRAFT框架以解决步级优势估计偏差问题 reinforcement learning large language model ✅
31 SciWalker: Synthesizing Scientific Coding Problems with Operator Graphs and Execution Feedback 提出SciWalker以合成科学编码问题并提升LLM训练数据质量 reinforcement learning large language model ✅
32 Structured Pose-Conditioned Flow Matching for Generative 5G CSI Augmentation 提出StructFlow-HPR以解决5G CSI数据收集成本高的问题 flow matching
33 Qwen-Planner-Agent: A Closed-Loop AI-for-AI Framework for Real-World Mobile Planner Agents 提出Qwen-Planner-Agent以解决移动规划代理的开发与迭代问题 reinforcement learning large language model
34 Learning to Ideate for Scientific Impact 提出基于引用影响的科学创意生成模型 reinforcement learning large language model
35 Just Ask Jev: Reinforcement Learning for Calibrated Decisions as a Zero-Shot Detector of AI Alignment Failures 提出Jev模型以解决AI对齐失败检测问题 reinforcement learning ✅
36 Right Choice of Classification Algorithms Based on Reinforcement Learning for Prediction of Non-Alcoholic Fatty Liver 提出基于强化学习的分类算法选择方法以预测非酒精性脂肪肝 reinforcement learning
37 CounterRoute: Self-Routed Reasoning via Hierarchical Counterfactual Credit Assignment 提出CounterRoute以解决推理模型效率低下问题 reinforcement learning instruction following
38 Understanding Perturbed Parameter Ensemble Sensitivities Using A Contrastive Learning Approach 提出对比学习方法以解析气候模拟中的参数敏感性问题 contrastive learning

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
39 AD-WM: Action-Discriminative World Models for Counterfactual Model Predictive Control 提出AD-WM以解决模型预测控制中的行动区分问题 MPC model predictive control world model ✅
40 AD-WM: Action-Discriminative World Models for Counterfactual Model Predictive Control 提出AD-WM以解决模型预测控制中的行动区分问题 MPC model predictive control world model ✅

⬅️ 返回 cs.AI 首页 · 🏠 返回主页