arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2101.04882cs.LGcs.AIcs.CVcs.RO
收藏

用于机器人操作中自动目标发现的非对称自游戏

Asymmetric self-play for automatic goal discovery in robotic manipulation

OpenAI
OpenAI OpenAI, Matthias Plappert, Raul Sampedro, Tao Xu, Ilge Akkaya, Vineet Kosaraju, Peter Welinder, Ruben D'Sa, Arthur Petron, Henrique P. d. O. Pinto, Alex … 展开作者

OpenAI OpenAI, Matthias Plappert, Raul Sampedro, Tao Xu, Ilge Akkaya, Vineet Kosaraju, Peter Welinder, Ruben D'Sa, Arthur Petron, Henrique P. d. O. Pinto, Alex Paino, Hyeonwoo Noh, Lilian Weng, Qiming Yuan, Casey Chu, Wojciech Zaremba

摘要:

我们训练一个单一的,目标条件的政策,可以解决许多机器人的操作任务,包括以前看不见的目标和对象的任务。我们依赖于非对称的自我游戏来进行目标发现,其中两个代理,Alice和Bob,进行游戏。爱丽丝被要求提出具有挑战性的目标,鲍勃的目标是解决这些问题。我们表明,这种方法可以发现高度多样化和复杂的目标,而无需任何人类先验知识。鲍勃可以只接受稀疏的奖励,因为爱丽丝和鲍勃之间的互动会产生一个自然的课程,鲍勃可以从爱丽丝的轨迹中学习,当爱丽丝被重新标记为目标条件演示时。最后,我们的方法可以扩展,从而产生一个单一的策略,可以推广到许多看不见的任务,如设置一个表,堆叠块,并解决简单的难题。学习策略的视频可在https://robotics-self-play.github.io上获得。

英文摘要:

We train a single, goal-conditioned policy that can solve many robotic manipulation tasks, including tasks with previously unseen goals and objects. We rely on asymmetric self-play for goal discovery, where two agents, Alice and Bob, play a game. Alice is asked to propose challenging goals and Bob aims to solve them. We show that this method can discover highly diverse and complex goals without any human priors. Bob can be trained with only sparse rewards, because the interaction between Alice and Bob results in a natural curriculum and Bob can learn from Alice's trajectory when relabeled as a goal-conditioned demonstration. Finally, our method scales, resulting in a single policy that can generalize to many unseen tasks such as setting a table, stacking blocks, and solving simple puzzles. Videos of a learned policy is available at https://robotics-self-play.github.io.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑