递归驾驭蒸馏：面向机器人操作的多智能体知识积累

中文摘要 AI 辅助

机器人学的核心目标之一是使机器人能够在不断变化的任务和环境中进行操作。视觉-语言-动作（VLA）模型提供了广泛的操作能力，但当执行需要诊断失败并调整行为时，这些模型可能会遇到困难。强大的智能体可以通过与这些策略的交互来发现有效的干预措施。我们提出递归驾驭蒸馏（Recursive Harness Distillation）方法，将这种经验作为可复用的指导在多个智能体之间积累。一个强大的智能体将其经验蒸馏成一份“操作手册”（playbook）供轻量级智能体使用，然后利用轻量级智能体的执行反馈递归地完善该手册。最终生成的手册使智能体能够在新的任务实例中复用积累的干预知识，而无需更新模型参数。在真实世界的操作任务中，该驾驭方法将成功率从37.3%提升至64.0%。在SimplerEnv Bridge基准上，配备手册的轻量级智能体达到了66.7%的成功率，而仅使用GR00T的基线成功率仅为41.7%，并且该轻量级智能体还超越了未配备手册的强智能体。同一手册也使强智能体受益，其成功率达到了79.2%。这些结果证明了驾驭蒸馏在机器人领域的可行性：干预经验可以被积累、通过执行进行精炼，并在多个智能体之间复用，从而提升操作性能。

英文摘要

A central goal in robotics is to enable manipulation across changing tasks and environments. Vision-language-action (VLA) models provide broad manipulation capabilities but can struggle when execution requires diagnosing failures and adapting behavior. Strong agents can discover effective interventions through interaction with these policies. We propose Recursive Harness Distillation to accumulate this experience as reusable guidance across agents. A strong agent distills its experience into a playbook for a light agent, then recursively refines the playbook using the light agent's execution feedback. The resulting playbook enables agents to reuse accumulated intervention knowledge in new task instances without updating model parameters. In real-world manipulation, the harness improves success from 37.3% to 64.0%. On SimplerEnv Bridge, the light agent with the playbook achieves 66.7% success, compared with 41.7% for the GR00T-only baseline, and outperforms the strong agent without a playbook. The same playbook also benefits the strong agent, which reaches 79.2% success. These results demonstrate the feasibility of harness distillation for robotics: intervention experience can be accumulated, refined through execution, and reused across agents to improve manipulation.

发表机构

• Seoul National University（首尔大学）
• Korea Advanced Institute of Science and Technology（韩国科学技术院）

机构由 AI 辅助整理，请以论文原文为准。