Hierarchical Learning with Hindsight (ICLR 19)

Agents learn to solve various continuous and discrete tasks using the Hierarchical Actor-Critic (HAC) and Hierarchical Q-Learning (HierQ) algorithms, respectively. These algorithms help agents break down long tasks into shorter tasks, enabling agents to learn tasks more quickly.

ICLR 2019 Paper: https://openreview.net/pdf?id=ryzECoAcY7
Blog Post: http://bigai.cs.brown.edu/2019/09/03/hac.html
GitHub Repo: https://github.com/andrew-j-levy/Hierarchical-Actor-Critc-HAC-