北京通用人工智能研究院BIGAI https://validator.w3.org/feed/docs/rss2.html 科技创新与产业创新深度融合，2026通用人工智能论坛成功召开 NarrativeLoom: Enhancing Creative Storytelling through Multi-Persona Collaborative Improvisation Read the Room: Video Social Reasoning with Mental-Physical Causal Chains SceneCOT: Eliciting Chain-of-Thought Reasoning in 3D Scenes G4Splat: Geometry-Guided Gaussian Splatting with Generative Prior Linking Process to Outcome: Conditonal Reward Modeling for LLM Reasoning Cross-Robot Behavior Adaptation through Intention Alignment 通研院研究成果登《Science Robotics》：首次实现异构机器人团队之间基于意图理解的自适应模仿与协作 产品能力领跑全球！通研院 TongAgents 登上多项国际智能体榜单 Aegis: Automated Error Generation and Identification for Multi-Agent Systems DexMove: Learning Tactile-Guided Non-Prehensile Manipulation with Dexterous Hands 智通未来 能者聚首：第十届北大智能青年论坛 Generating Objects with Part-Articulation from a Single Image Towards Brigding the Gap Between Large-scale Pretraining and Efficient Finetuning for Humanoid Control STVG-R1: Incentivizing Instance-Level Reasoning and Grounding in Videos via Reinforcement Learning MVR:Multi-view Video Reward Shaping for Reinforcement Learning MILR: Improving Multimodal Image Generation via Test-Time Latent Reasoning When Large Multimodal Models Confront Evolving Knowledge: Challenges and Explorations Learning What Matters Now: Dynamic Preference Inference under Contextual Shifts RuleReasoner: Reinforced Rule-based Reasoning via Domain-aware Dynamic Sampling PoliCon: Evaluating LLMs on Achieving Diverse Political Consensus Objectives Proposing and solving olympiad geometry with guided tree search TongUI: Building Generalized GUI Agents by Learning from Multimodal Web Tutorials A VR-Based Robotic Teleoperation System With Haptic Feedback and Adaptive Collision Avoidance 3D Scene Change Modeling With Consistent Multi-View Aggregation 通研院、宇树提出首个攻克泛化壁垒的通用运动控制框架OmniXtreme：让机器人解锁“托马斯全旋”等极限动作 M3Bench: Benchmarking Whole-Body Motion Generation for Mobile Manipulation in 3D Scenes 通研院AI“全能物理空间训练场”TongSIM开源上线，两天登上Huggingface热度趋势榜单第一名 2026人工智能教育创新发展论坛暨海淀街道首届产学结合高校通用人工智能大赛颁奖仪式成功举办 通研院、北大突破性成果登《自然·机器智能》：全球首个具备自主出题与解题能力的奥数几何AI模型 教育部召开代表委员座谈会，朱松纯提出智能时代“何以为人”的新内涵 通研院发布行业智能体基准，重新定义“数字员工”上岗标准 共筑“海淀AI教育生态圈”，通研院教育科技成果亮相海淀人工智能教育庙会 Integration of Robot and Scene Kinematics for Sequential Mobile Manipulation Planning PR2: A Physics- and Photo-realistic Humanoid Testbed with Pilot Study in Competition Adaptive Preference Optimization with Uncertainty-aware Utility Anchor Reinforced Query Reasoners for Reasoning-intensive Retrieval Tasks Mind the Gap: The Divergence Between Human and LLM-Generated Tasks Understanding and Leveraging the Expert Specialization of Context Faithfulness in Mixture-of-Experts LLMs Enhancing LLM-Based Social Bot via an Adversarial Learning Framework From Objects to Anywhere: A Holistic Benchmark for Multi-level Visual Grounding in 3D Scenes Iterative Tool Usage Exploration for Multimodal Agents via Step-wise Preference Tuning All-in-one 3D Scene Synthesis with an Extensible and Self-Reflective Agent Reasoning with Exploration: An Entropy Perspective Taccel: Scaling Up Vision-based Tactile Robotics via High-performance GPU Simulation Absolute Zero: Reinforced Self-play Reasoning with Zero Data ADAPT: Adaptive Decentralized Architecture with Perception-aligned Training for Structural Generalization in Multi-Agent RL World Models Should Prioritize the Unification of Physical and Social Dynamics Social World Model-Augmented Mechanism Design Policy Learning 主页