Online policy learning directly in the physical world is a promising yet challenging direction for embodied intelligence. Unlike simulation, real-world systems cannot be arbitrarily accelerated, cheaply reset, or massively replicated, suggesting that real-world policy learning is not merely an algorithmic problem, but inherently a systems problem. We present \user, a \underline{U}nified and extensible \underline{S}yst\underline{E}m for online real-world policy lea\underline{R}ning. On the systems side, \user introduces a hardware abstraction layer for unified robot management and an adaptive communication plane that enables efficient cloud-edge training. On the learning side, \user adopts a fully asynchronous training framework, designs a persistent and cache-aware replay buffer, and provides extensible abstractions for rewards, algorithms, and policies. Experiments in both simulation and the real world demonstrate that \user supports multi-robot coordination, heterogeneous manipulators, cloud-edge training with large models, and long-running asynchronous training. Together, these capabilities establish \user as a unified and extensible systems foundation for online real-world policy learning.
Enabling Extensible Embodied Capabilities with Tools
Huazhong University of Science and Technology(华中科技大学);Hebei University of Technology(河北工业大学);Tianjin University(天津大学);Lehigh University(莱特大学);College of AI, Tsinghua University(清华大学人工智能学院)
Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots
Southeast University(东南大学);Nanjing University(南京大学);Microsoft Research(微软研究院);Institute for AI Industry Research (AIR), Tsinghua University(清华大学人工智能产业研究院(AIR))
Do Vision-Language-Action Models Mean What They Say? On the Role of Faithfulness in Embodied Reasoning
Stanford University(斯坦福大学);Politecnico di Milano(米兰理工大学);KTH Royal Institute of Technology(皇家理工学院);Italian Institute of Artificial Intelligence (AI4I)(意大利人工智能研究所);NVIDIA Research(英伟达研究院)
EmbodiedClaw: Conversational Workflow Execution for Embodied AI Development
School of Cyber Science and Engineering, Huazhong University of Science and Technology(华中科技大学信息科学与工程学院);School of Mechanical Science and Engineering, Huazhong University of Science and Technology(华中科技大学机械科学与工程学院);School of Information and Software Engineering, University of Electronic Science and Technology of China(电子科技大学信息与软件工程学院);Lehigh University(莱斯大学);College of AI, Tsinghua University(清华大学人工智能学院)
EmbodiedGovBench: A Benchmark for Governance, Recovery, and Upgrade Safety in Embodied Agent Systems
School of Software, Harbin Institute of Technology(哈尔滨工业大学软件学院);School of Computer Science and Technology, Harbin Institute of Technology(哈尔滨工业大学计算机科学与技术学院);School of Mathematical and Computer Sciences, Heriot-Watt University, Malaysia Campus(赫瑞-瓦特大学马来西亚校区数学与计算机科学学院);School of Future Science and Engineering, Soochow University(苏州大学未来科学与工程学院)
Embodied AI: Emerging Risks and Opportunities for Policy Action
Centre for the Governance of AI(治理人工智能中心);French Center for AI Safety (CeSIA)(法国人工智能安全中心);Carnegie Mellon University(卡内基梅隆大学);University of Oxford(牛津大学);University of Bologna(博洛尼亚大学);Tony Blair Institute for Global Change(托尼·布莱尔全球变革研究所);Yale University(耶鲁大学);Yale Digital Ethics Center(耶鲁数字伦理中心)
Towards Deploying VLA without Fine-Tuning: Plug-and-Play Inference-Time VLA Policy Steering via Embodied Evolutionary Diffusion
Collaborative and Versatile Robots (CLOVER) Laboratory, T-Stone Robotics Institute, The Chinese University of Hong Kong, Hong Kong(协作与多功能机器人实验室,T-Stone机器人研究所,香港中文大学,香港);Φ \Phi -Institute for Physical Human Intelligence(物理人机智能研究所);Center for Embodied Artificial Intelligence and Computer Vision, Shenzhen Loop Area Institute, Shenzhen, China(具身人工智能与计算机视觉中心,深圳环园研究院,深圳,中国);Department of Advanced Robotics, Istituto Italiano di Tecnologia, Genoa, Italy(先进机器人系,意大利理工学院,热那亚,意大利)
EmbodiedSkills: World Action Model-augmented Cascaded Skills for Vision-Language-Action Agent
College of Computer Science and Technology, Zhejiang University(浙江大学计算机科学与技术学院);Nanjing University of Aeronautics and Astronautics(南京航空航天大学);Cornell University(康奈尔大学);Universal Ubiquitous AI Co., Ltd.(Universal Ubiquitous AI有限公司);Hangzhou DEEP Robotics Technology Co., Ltd.(杭州深度机器人科技有限公司);National University of Singapore(新加坡国立大学)
LAPT-Scale: Scaling Embodied Policies with Latent Action Pre-Training on Human Videos
Joy Future Academy, JD(京东探索研究院);School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen(哈尔滨工业大学(深圳)计算机科学与技术学院);Tsinghua SIGS, Tsinghua University(清华大学深圳国际研究生院);Artificial Intelligence, The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州)人工智能学院)
Look, Zoom, Understand: The Robotic Eyeball for Embodied Perception
School of Artificial Intelligence, Shanghai Jiao Tong University(上海交通大学人工智能学院);Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);Dalian University of Technology(大连理工大学)
FluxVLA Engine: A One-Stop VLA Engineering Platform for Embodied Intelligence
LimX Dynamics;Nankai University(南开大学);The University of Hong Kong(香港大学);University of Electronic Science and Technology of China(电子科技大学);Beijing Institute of Technology(北京理工大学);Xidian University(西安电子科技大学);Zhejiang University(浙江大学);Xi’an Jiaotong University(西安交通大学)
VAG: Dual-Stream Video-Action Generation for Embodied Data Synthesis
GigaAI(极佳科技);Zhejiang University(浙江大学);Peking University(北京大学);Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学)
Speedup Patch: Learning a Plug-and-Play Policy to Accelerate Embodied Manipulation
National Key Laboratory for Novel Software Technology, Nanjing University, China;School of Artificial Intelligence, Nanjing University, China;Mila-Quebec AI Institute \& Université de Montréal, Canada
Harnessing Embodied Agents: Runtime Governance for Policy-Constrained Execution
School of Software, Harbin Institute of Technology(哈尔滨工业大学软件学院);School of Computer Science and Technology, Harbin Institute of Technology(哈尔滨工业大学计算机科学与技术学院);School of Mathematical and Computer Sciences, Heriot-Watt University, Malaysia Campus(赫瑞-沃德大学马来西亚校区数学与计算机科学学院);School of Future Science and Engineering, Soochow University(苏州大学未来科学与工程学院);Fraunhofer Institute for Applied Information Technology(弗劳恩霍夫应用信息技术研究所)