Causality-driven Hierarchical Structure Discovery for Reinforcement Learning
This addresses the problem of low exploration efficiency in HRL for complex environments, offering a novel approach but likely incremental as it builds on existing HRL methods.
The paper tackles the challenge of automatically discovering high-quality hierarchical structures in hierarchical reinforcement learning (HRL) for tasks with sparse rewards, proposing CDHRL, a causality-driven framework that replaces randomness-driven exploration. The results show that CDHRL significantly boosts exploration efficiency in complex environments like 2D-Minecraft and Eden.
Hierarchical reinforcement learning (HRL) effectively improves agents' exploration efficiency on tasks with sparse reward, with the guide of high-quality hierarchical structures (e.g., subgoals or options). However, how to automatically discover high-quality hierarchical structures is still a great challenge. Previous HRL methods can hardly discover the hierarchical structures in complex environments due to the low exploration efficiency by exploiting the randomness-driven exploration paradigm. To address this issue, we propose CDHRL, a causality-driven hierarchical reinforcement learning framework, leveraging a causality-driven discovery instead of a randomness-driven exploration to effectively build high-quality hierarchical structures in complicated environments. The key insight is that the causalities among environment variables are naturally fit for modeling reachable subgoals and their dependencies and can perfectly guide to build high-quality hierarchical structures. The results in two complex environments, 2D-Minecraft and Eden, show that CDHRL significantly boosts exploration efficiency with the causality-driven paradigm.