arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1903.00636cs.RO
收藏

基于人类对抗博弈的机器人学习

Robot Learning via Human Adversarial Games

University of Southern California(南加州大学);Carnegie Mellon University(卡内基梅隆大学)

Jiali Duan, Qian Wang, Lerrel Pinto, C. -C. Jay Kuo, Stefanos Nikolaidis

摘要:

机器人领域的许多工作都集中在提高学习过程效率的“人在回路”学习技术上。然而,这些算法已经做出了强烈的假设,即协助机器人的合作人类监督者。在现实中,人类观察者也倾向于以对抗的方式对待部署的机器人系统。我们表明,这实际上可以通过提出一个物理框架来提高学习模型的鲁棒性,该物理框架利用人类对手施加的扰动,引导机器人走向更鲁棒的模型。在一个操作任务中,我们表明,抓成功显着提高时,机器人与人类对手的训练相比,在自我监督的方式训练。

英文摘要:

Much work in robotics has focused on "human-in-the-loop" learning techniques that improve the efficiency of the learning process. However, these algorithms have made the strong assumption of a cooperating human supervisor that assists the robot. In reality, human observers tend to also act in an adversarial manner towards deployed robotic systems. We show that this can in fact improve the robustness of the learned models by proposing a physical framework that leverages perturbations applied by a human adversary, guiding the robot towards more robust models. In a manipulation task, we show that grasping success improves significantly when the robot trains with a human adversary as compared to training in a self-supervised manner.

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。

↑