Relative Variational Intrinsic Control

TL;DR

This paper introduces Relative Variational Intrinsic Control (RVIC), a new skill learning method that encourages agents to develop diverse, environment-influencing behaviors, improving hierarchical reinforcement learning performance.

Contribution

The paper proposes RVIC, a novel objective that promotes meaningful skill diversity based on how skills alter the agent's environmental relationship, addressing limitations of existing mutual information approaches.

Findings

RVIC skills are more useful in hierarchical RL tasks.

Qualitative analysis shows diverse and meaningful skill behaviors.

RVIC outperforms existing skill discovery methods.

Abstract

In the absence of external rewards, agents can still learn useful behaviors by identifying and mastering a set of diverse skills within their environment. Existing skill learning methods use mutual information objectives to incentivize each skill to be diverse and distinguishable from the rest. However, if care is not taken to constrain the ways in which the skills are diverse, trivially diverse skill sets can arise. To ensure useful skill diversity, we propose a novel skill learning objective, Relative Variational Intrinsic Control (RVIC), which incentivizes learning skills that are distinguishable in how they change the agent's relationship to its environment. The resulting set of skills tiles the space of affordances available to the agent. We qualitatively analyze skill behaviors on multiple environments and show how RVIC skills are more useful than skills discovered by existing methods when used in hierarchical reinforcement learning.