Diversity is All You Need: Learning Skills without a Reward Function

TL;DR

This paper introduces DIAYN, a method for unsupervised skill learning that maximizes diversity without reward signals, enabling agents to develop useful behaviors and improve downstream task performance.

Contribution

The paper presents DIAYN, a novel unsupervised learning approach that discovers diverse skills through an information-theoretic objective, reducing reliance on reward functions in reinforcement learning.

Findings

Unsupervised skills include walking and jumping behaviors.

Pretrained skills improve data efficiency in downstream tasks.

Skills can be combined hierarchically to solve complex problems.

Abstract

Intelligent creatures can explore their environments and learn useful skills without supervision. In this paper, we propose DIAYN ('Diversity is All You Need'), a method for learning useful skills without a reward function. Our proposed method learns skills by maximizing an information theoretic objective using a maximum entropy policy. On a variety of simulated robotic tasks, we show that this simple objective results in the unsupervised emergence of diverse skills, such as walking and jumping. In a number of reinforcement learning benchmark environments, our method is able to learn a skill that solves the benchmark task despite never receiving the true task reward. We show how pretrained skills can provide a good parameter initialization for downstream tasks, and can be composed hierarchically to solve complex, sparse reward tasks. Our results suggest that unsupervised discovery of skills can serve as an effective pretraining mechanism for overcoming challenges of exploration and data efficiency in reinforcement learning.