外观
论文arXiv 2610.0752723 天前
用于高效强化学习的任务空间模仿引导
原文链接：Task-Space Imitation Guidance for Efficient Reinforcement Learning（arXiv 2610.07527 · 发布日期：2026-10-05）
Task-Space Imitation Guidance for Efficient Reinforcement Learning
TIGER面向稀疏奖励桌面机器人操作，用任务空间模仿引导构造奖励与预训练，并在仿真和真机实验中验证。
核心要点
- 针对稀疏奖励下的桌面机器人操作，提升早期在线强化学习的样本效率与安全性。
- 将动作分块模仿策略作为局部任务空间进展估计器，而非可执行控制器或动作先验。
- 把预测动作块映射为短时末端执行器参考，给予密集进展奖励，同时保留稀疏环境奖励主导。
- 仿真与真机实验显示其提升早期样本效率、减少安全违规，并达到或提升最终成功率；提供视频和概览链接。
验证范围与局限：验证范围限于摘要所述仿真和真实机器人上的桌面操作评估任务。
英文摘要 (Abstract)
We introduce Task-Space Imitation Guidance for Efficient Reinforcement Learning (TIGER), a reward-construction and pretraining framework for sparse-reward tabletop robotic manipulation. TIGER treats an action-chunked imitation policy not as an executable controller or action prior, but as a local task-space progress estimator: predicted action chunks are converted, using controller-aware action-to-motion mapping, into short-horizon end-effector references, and the RL agent receives dense progress rewards toward these references while the sparse environment reward remains the dominant objective. During pretraining, TIGER uses imitation-guided look-ahead signals to relax conservative value penalties for actions predicted to make task-space progress, reducing off-manifold exploration during early online RL. Across simulation and real-robot experiments, TIGER improves early sample efficiency and reduces measured safety violations while matching or improving final success rates relative to prior RL and IL-RL baselines on the evaluated tasks. Videos and an overview of this work are available at https://sanctuary.ai/news/tiger/.
作者：Salar Asayesh, Hossein Darani, Todd Cao, Evgeny Andriash, Mani Ranjbar
来源
本页事实对应的公开页面，以原文为准。