一人称視点の一回デモンストレーションからロボットマニピュレーションを学習するO2A
O2A: One-shot Observational learning with Action vectors
第三者視点の一回のデモ動画から、事前学習した3D-CNNで行動ベクトルを抽出し、ロボットの試行との距離を報酬として強化学習することで、視点や物体・背景・マニピュレータ形態の変化に頑健なマニピュレーション学習を実現した。
著者: Leo Pauly, Wisdom C. Agboh, David C. Hogg, Raul Fuentes
分類: cs.RO, cs.LG
原文アブストラクト
We present O2A, a novel method for learning to perform robotic manipulation tasks from a single (one-shot) third-person demonstration video. To our knowledge, it is the first time this has been done for a single demonstration. The key novelty lies in pre-training a feature extractor for creating a perceptual representation for actions that we call 'action vectors'. The action vectors are extracted using a 3D-CNN model pre-trained as an action classifier on a generic action dataset. The distance between the action vectors from the observed third-person demonstration and trial robot executions is used as a reward for reinforcement learning of the demonstrated task. We report on experiments in simulation and on a real robot, with changes in viewpoint of observation, properties of the objects involved, scene background and morphology of the manipulator between the demonstration and the learning domains. O2A outperforms baseline approaches under different domain shifts and has comparable performance with an oracle (that uses an ideal reward function).
