Learning Multi-Level Hierarchies with Hindsight

TL;DR

This paper introduces Hierarchical Actor-Critic (HAC), a new framework for hierarchical reinforcement learning that enables stable, parallel learning of multiple policy levels, significantly improving learning speed and scalability in complex tasks.

Contribution

HAC allows independent, parallel training of hierarchy levels by assuming lower levels are optimal, overcoming instability issues in hierarchical RL.

Findings

HAC accelerates learning compared to other methods.

First to learn 3-level hierarchies in continuous spaces.

Effective in grid world and robotics domains.

Abstract

Hierarchical agents have the potential to solve sequential decision making tasks with greater sample efficiency than their non-hierarchical counterparts because hierarchical agents can break down tasks into sets of subtasks that only require short sequences of decisions. In order to realize this potential of faster learning, hierarchical agents need to be able to learn their multiple levels of policies in parallel so these simpler subproblems can be solved simultaneously. Yet, learning multiple levels of policies in parallel is hard because it is inherently unstable: changes in a policy at one level of the hierarchy may cause changes in the transition and reward functions at higher levels in the hierarchy, making it difficult to jointly learn multiple levels of policies. In this paper, we introduce a new Hierarchical Reinforcement Learning (HRL) framework, Hierarchical Actor-Critic (HAC), that can overcome the instability issues that arise when agents try to jointly learn multiple levels of policies. The main idea behind HAC is to train each level of the hierarchy independently of the lower levels by training each level as if the lower level policies are already optimal. We demonstrate experimentally in both grid world and simulated robotics domains that our approach can significantly accelerate learning relative to other non-hierarchical and hierarchical methods. Indeed, our framework is the first to successfully learn 3-level hierarchies in parallel in tasks with continuous state and action spaces.