Augmenting Policy Learning with Routines Discovered from a Single Demonstration
This work addresses the problem of accelerating skill learning for AI agents by leveraging prior knowledge from minimal data, which is an incremental improvement for reinforcement learning and imitation learning practitioners.
This paper introduces Routine-Augmented Policy Learning (RAPL), a method that extracts routines from a single demonstration by identifying grammar over action trajectories. RAPL then uses these routines to augment policy learning at both primitive and routine levels, improving state-of-the-art imitation learning (SQIL) and reinforcement learning (A2C) on Atari games.
Humans can abstract prior knowledge from very little data and use it to boost skill learning. In this paper, we propose routine-augmented policy learning (RAPL), which discovers routines composed of primitive actions from a single demonstration and uses discovered routines to augment policy learning. To discover routines from the demonstration, we first abstract routine candidates by identifying grammar over the demonstrated action trajectory. Then, the best routines measured by length and frequency are selected to form a routine library. We propose to learn policy simultaneously at primitive-level and routine-level with discovered routines, leveraging the temporal structure of routines. Our approach enables imitating expert behavior at multiple temporal scales for imitation learning and promotes reinforcement learning exploration. Extensive experiments on Atari games demonstrate that RAPL improves the state-of-the-art imitation learning method SQIL and reinforcement learning method A2C. Further, we show that discovered routines can generalize to unseen levels and difficulties on the CoinRun benchmark.