cs.AI（2026-09-25）

📊 共 38 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) (24 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (11 🔗1) 支柱一：机器人控制 (Robot Control) (2 🔗1) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (24 篇)

#题目一句话要点标签🔗⭐
1 Can Linguistic Reasoning Vectors Enhance Multimodal Reasoning Ability? 提出LIFT方法以恢复多模态模型的语言推理能力 large language model multimodal
2 Training Graph Foundation Models on The Web Graph 提出Acacia模型以解决图基础模型的训练局限性 foundation model
3 EXAONE Demand 1.0: A Time Series Foundation Model for Demand Forecasting 提出EXAONE Demand以解决需求预测中的数据稀缺问题 foundation model
4 Neural State Prediction: Obstructing Shortcut Learning in EEG Foundation Models 提出神经状态预测以解决EEG模型中的快捷学习问题 foundation model
5 BAT-CLIP: Trimodal Alignment of Brain, Audio and Text 提出BAT-CLIP以解决多模态脑-音频-文本对齐问题 foundation model multimodal
6 Prompt Minimization: Reducing Input Redundancy Without Sacrificing Output Fidelity 提出提示最小化方法以减少输入冗余而不牺牲输出保真度 large language model
7 Game Arena: Strategic LLM Evaluation in Competitive Environments 提出Kaggle游戏竞技场以评估大型语言模型的策略能力 large language model
8 Segment-Level Agentic Topic Modeling for Improved Data Exploration and Resource Efficiency 提出SeLATM以解决LLM主题建模中的资源消耗问题 large language model
9 Towards Mitigating Fabricated Consensus: The Active Provenance Gate for Multi-Agent Debate Synthesis 提出主动来源门以解决多智能体辩论合成中的虚假共识问题 large language model
10 Intent2Tc: Automated Intent-to-Traffic Control Translation with Language Models 提出Intent2Tc以解决高层服务意图与网络配置间的翻译问题 large language model
11 Mutable Transcripts: Mitigating Context Pollution through Editable Conversation State 提出可编辑对话状态以缓解上下文污染问题 large language model ✅
12 Beyond Approved Actions: Runtime Validation of Persistent Outcomes in Agent Workflows 提出EffectMatch以解决代理工作流中的持久结果验证问题 large language model
13 DIAL: Position-Debiased LLM Judges with Adaptive Human Preference Calibration 提出DIAL框架以解决LLM评判中的位置偏差问题 large language model
14 Rethinking Data Quality for AI-Driven Systems: Evidence from Practitioner Interviews 提出数据质量管理框架以应对AI驱动系统中的信任问题 foundation model
15 OmouAI: Argumentative Human-AI Policy Deliberation with Simulated Personas 提出OmouAI以解决高风险环境中的人机政策辩论问题 large language model
16 Externalized CPDAG Summaries Improve LLM Causal Deduction 提出结构化思维以提升大语言模型的因果推断能力 chain-of-thought
17 Cheap, open agents make LLM pollution harder to mitigate 提出开放代理以应对大型语言模型污染问题 large language model
18 Same Text, Different Numbers: The Divergence of LLM-Based Measures 探讨LLM模型选择对文本测量结果的影响 large language model
19 MoMHa: Multi-Objective Optimization of LLM Harnesses over Accuracy, Safety, and Tokens 提出MoMHa以解决大型语言模型的多目标优化问题 large language model
20 FTB Graph: Determining and Validating First-token Broadcasters and Language-Identity Head Circuits in Multilingual Language Models 提出FTB图以解析多语言模型中的首个令牌广播机制 large language model
21 LogicTree-RAG: Logic Tree-guided Retrieval-Augmented Generation for Long-form Patent Drafting 提出LogicTree-RAG以解决长篇专利撰写中的逻辑结构与技术推理问题 large language model
22 Why Jailbreaks Succeed in Diffusion Language Models: An Energy Landscape Analysis 通过能量景观分析提出新方法以解释扩散语言模型的越狱成功原因 large language model
23 Evaluating Real-Time Voice Agents: From Component Quality to Grounded Outcomes 提出TRG标准以评估实时语音代理的性能与效果 foundation model
24 A Framework for Identifying, Categorizing, and Explaining Bias in AI-Generated Code 提出框架以识别、分类和解释AI生成代码中的偏见 large language model

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

#题目一句话要点标签🔗⭐
25 Self-Play Search Distillation for Large Language Model Reasoning 提出自我对弈搜索蒸馏以提升大语言模型推理能力 distillation large language model
26 Monitor Jailbreaking: Evading Chain-of-Thought Monitoring Without Encoded Reasoning 提出监控越狱方法以规避链式思维监控问题 reinforcement learning chain-of-thought
27 TISD: On-Policy Self-Distillation with Trajectory Intervention 提出TISD以解决现有自蒸馏方法的训练瓶颈问题 teacher-student distillation
28 A Flow Matching Framework for Neural Representational Dissimilarity 提出流匹配框架以统一神经表征差异性度量 flow matching
29 Compress What You See, Not What You Say: Anchored Context Distillation for Latent-Observation Software Engineering Agents 提出LOHA与ACD以解决软件工程代理的上下文压缩问题 distillation
30 MA-WAM: Multi-Agent World-Action Model for Test-Time Planning 提出MA-WAM以解决多智能体协作任务中的规划问题 reinforcement learning world model world models
31 Samples, Sources, Space: Decomposing Data Scale in Spatially Structured Representation Learning of Human Brain Microarchitecture 提出数据分配方法以优化人脑微观结构的表示学习 representation learning
32 Momentum-Guided Federated Split Distillation for Personalized Temporal Edge Intelligence 提出动量引导的联邦分割蒸馏框架以实现个性化边缘智能 distillation
33 JevAdvBench: A Benchmark and Black-Box Attacks for Reinforcement Learning for Calibrated Decisions Models 提出JevAdvBench以评估强化学习决策模型的鲁棒性 reinforcement learning ✅
34 AtomWorld-Mem: Memory-Restored World States for Long-Horizon Atomistic Evolution 提出AtomWorld-Mem以解决长时间尺度原子演化中的记忆恢复问题 world model world models
35 From S3Q Theory to Implementation: Towards an Architecture for Machine Qualia 提出五层架构以实现S3Q意识理论的计算模型 world model world models

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
36 ORCA: Evaluating LLMs on Data Science Code Translation 提出ORCA基准以评估LLMs在数据科学代码翻译中的表现 manipulation large language model
37 SciHorizon-eLab: An Agentic Protocol-to-Task Compiler for Scalable Benchmarking of Scientific Embodied Agents 提出SciHorizon-eLab以解决科学实验自动化评估问题 manipulation ✅

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

#题目一句话要点标签🔗⭐
38 Subject-Invariant Cross-Modal Decoding of Perceived Speech from Brain Recordings 提出SICMD方法以解决跨主体感知语音解码问题 spatiotemporal

⬅️ 返回 cs.AI 首页 · 🏠 返回主页