Learning Temporal Abstraction with Information-theoretic Constraints for Hierarchical Reinforcement Learning

Applying reinforcement learning (RL) to real-world problems will require reasoning about action-reward correlation over long time horizons. Hierarchical reinforcement learning (HRL) methods handle this by dividing the task into hierarchies, often with hand-tuned network structure or pre-defined subgoals. We propose a novel HRL framework TAIC, which learns the temporal abstraction from past experience or expert demonstrations without task-specific knowledge. We formulate the temporal abstraction problem as learning latent representations of action sequences and present a novel approach of regularizing the latent space by adding information-theoretic constraints. Specifically, we maximize the mutual information between the latent variables and the state changes. A visualization of the latent space demonstrates that our algorithm learns an effective abstraction of the long action sequences. The learned abstraction allows us to learn new tasks on higher level more efficiently. We convey a significant speedup in convergence over benchmark learning problems. These results demonstrate that learning temporal abstractions is an effective technique in increasing the convergence rate and sample efficiency of RL algorithms.