Align-RUDDER: Learning From Few Demonstrations by Reward Redistribution

TL;DR

Align-RUDDER enhances reinforcement learning from limited demonstrations by using sequence alignment for reward redistribution, significantly improving learning efficiency in complex tasks with sparse rewards.

Contribution

It introduces a novel reward redistribution method based on multiple sequence alignment, enabling effective learning from few demonstrations in complex hierarchical tasks.

Findings

Outperforms existing methods on artificial tasks with delayed rewards.

Successfully mines a diamond in Minecraft with few demonstrations.

Drastically improves learning efficiency in sparse reward scenarios.

Abstract

Reinforcement learning algorithms require many samples when solving complex hierarchical tasks with sparse and delayed rewards. For such complex tasks, the recently proposed RUDDER uses reward redistribution to leverage steps in the Q-function that are associated with accomplishing sub-tasks. However, often only few episodes with high rewards are available as demonstrations since current exploration strategies cannot discover them in reasonable time. In this work, we introduce Align-RUDDER, which utilizes a profile model for reward redistribution that is obtained from multiple sequence alignment of demonstrations. Consequently, Align-RUDDER employs reward redistribution effectively and, thereby, drastically improves learning on few demonstrations. Align-RUDDER outperforms competitors on complex artificial tasks with delayed rewards and few demonstrations. On the Minecraft ObtainDiamond task, Align-RUDDER is able to mine a diamond, though not frequently. Code is available at https://github.com/ml-jku/align-rudder. YouTube: https://youtu.be/HO-_8ZUl-UY