English

Hierarchical Empowerment: Towards Tractable Empowerment-Based Skill Learning

Machine Learning 2023-10-05 v2 Artificial Intelligence Robotics

Abstract

General purpose agents will require large repertoires of skills. Empowerment -- the maximum mutual information between skills and states -- provides a pathway for learning large collections of distinct skills, but mutual information is difficult to optimize. We introduce a new framework, Hierarchical Empowerment, that makes computing empowerment more tractable by integrating concepts from Goal-Conditioned Hierarchical Reinforcement Learning. Our framework makes two specific contributions. First, we introduce a new variational lower bound on mutual information that can be used to compute empowerment over short horizons. Second, we introduce a hierarchical architecture for computing empowerment over exponentially longer time scales. We verify the contributions of the framework in a series of simulated robotics tasks. In a popular ant navigation domain, our four level agents are able to learn skills that cover a surface area over two orders of magnitude larger than prior work.

Keywords

Cite

@article{arxiv.2307.02728,
  title  = {Hierarchical Empowerment: Towards Tractable Empowerment-Based Skill Learning},
  author = {Andrew Levy and Sreehari Rammohan and Alessandro Allievi and Scott Niekum and George Konidaris},
  journal= {arXiv preprint arXiv:2307.02728},
  year   = {2023}
}

Comments

Additional baseline comparisons

R2 v1 2026-06-28T11:23:18.985Z