Few-Shot Imitation Learning 通过查询未标记的数据集进行行为检索（英文版）

英文标题：Behavior Retrieval: Few-Shot Imitation Learning by Querying Unlabeled Datasets中文摘要：本研究通过利用少量下游专家数据有针对性地获取有关未标记数据集的信息，设计了一种新颖的机器人学习方法以有效地学习视觉运动技能。该 方法效果显著，在通过视频实景测试和实际机器人操作中超越了以往更为复杂的目标设定方法 20% 的表现英文摘要：Enabling robots to learn novel visuomotor skills in a data-efficient mannerremains an unsolved problem with myriad challenges. A popular paradigm fortackling this problem is through leveraging large unlabeled datasets that havemany behaviors in them and then adapting a policy to a specific ta