Do's and Don'ts: Learning Desirable Skills with Instruction Videos

TL;DR

This paper introduces DoDont, a two-stage instruction-based skill discovery method that uses videos to guide learning towards desirable behaviors and away from unsafe or undesirable ones in continuous control tasks.

Contribution

The paper proposes a novel instruction-based approach that leverages videos to improve unsupervised skill discovery, emphasizing safety and desirability in learned behaviors.

Findings

Effective learning of desirable behaviors with less than 8 videos

Avoidance of unsafe behaviors like tripping and rolling

Successful application to complex control tasks

Abstract

Unsupervised skill discovery is a learning paradigm that aims to acquire diverse behaviors without explicit rewards. However, it faces challenges in learning complex behaviors and often leads to learning unsafe or undesirable behaviors. For instance, in various continuous control tasks, current unsupervised skill discovery methods succeed in learning basic locomotions like standing but struggle with learning more complex movements such as walking and running. Moreover, they may acquire unsafe behaviors like tripping and rolling or navigate to undesirable locations such as pitfalls or hazardous areas. In response, we present DoDont (Do's and Don'ts), an instruction-based skill discovery algorithm composed of two stages. First, in an instruction learning stage, DoDont leverages action-free instruction videos to train an instruction network to distinguish desirable transitions from undesirable ones. Then, in the skill learning stage, the instruction network adjusts the reward function of the skill discovery algorithm to weight the desired behaviors. Specifically, we integrate the instruction network into a distance-maximizing skill discovery algorithm, where the instruction network serves as the distance function. Empirically, with less than 8 instruction videos, DoDont effectively learns desirable behaviors and avoids undesirable ones across complex continuous control tasks. Code and videos are available at https://mynsng.github.io/dodont/