Runhan Huang 介绍 Equilibrium Energy Control(EqC),让机器人控制策略在生成动作的同时判断自身能力边界。方法采用能量式动作生成器,通过动作细化过程的几何特征监测自身能力,使策略能够决定何时请求帮助。

EqC 将 Energy-Gated DAgger 与 Astra 结合成改进循环。策略发现能力缺口后,请求智能体提供恢复动作,再从这些干预中学习,逐步扩大能够自主完成的任务范围。

据作者报告,实机衣物折叠任务在仅 6 次 Astra 生成的恢复后,自主成功率从 0% 提升至 100%。实机积木堆叠成功率从 0% 提升至 96.7%,碎屑清理任务成功率为 100%。

作者还报告,将 EqC 接入 π₀ 后,RoboTwin 2.0 Hard 上的成功率从 18% 提升至 33%,LIBERO-Plus 从 53.6% 提升至 65.8%。代码已开源,论文与项目演示同步提供。

原帖留影

What should control policies and VLAs look like in the era of agentic robotics? Policies (e.g., VLAs and WAMs) should know their capability limits, call an agent for help, and turn every intervention into greater autonomy. Introducing Equilibrium Energy Control (EqC): control with introspection, built for self-improving robots. 🤖 🧠 One energy landscape. Both action and introspection. Using an energy-based action generator, EqC generates actions while monitoring its own competence through the geometry of action refinement. The policy itself knows when to ask for help. 🔄 Energy-Gated DAgger × Astra: an agent-powered self-improvement loop. Detect a competence gap → request agent recovery → learn from it → expand autonomy. A concrete recipe for RSI in VLAs. More effective than human judged intervention. • Real-world garment folding: 0% → 100% autonomous success after just 6 Astra-generated recoveries. • Real-world block stacking: 0% → 96.7%. • Real-world debris cleaning: 100% success. 🚀 Stronger control across four benchmarks. EqC boosts π₀ on RoboTwin 2.0 Hard (18% → 33%) and LIBERO-Plus (53.6% → 65.8%) in a plug-and-play manner! Accepted as a Spotlight at the Continually Self-Improving Robots (CSIR) Workshop @ CoRL 2026! Come see the poster—and join for my spotlight talk: 🗓 Nov 12, 9:25–9:30 AM CT 📍 Room 202, JW Marriott Austin Code is open source! 📄 Paper: https://arxiv.org/abs/2610.18790 🌐 Project & demos: https://equilibrium-energy-control.github.io 💻 GitHub: https://github.com/hrh6666/EqC #RobotLearning #AI #LLM #Agents #RSI #corl2026 #corl #VLA #Control #Robotics #EBM
@RunhanH在 X 查看原帖发布于 2026/11/12 10:55(北京时间)