Hierarchical Deep Reinforcement Learning: Integrating Temporal Abstraction and Intrinsic Motivation

TL;DR

This paper introduces hierarchical-DQN (h-DQN), a deep reinforcement learning framework that combines hierarchical value functions with intrinsic motivation to improve exploration and learning in environments with sparse feedback.

Contribution

The paper presents a novel hierarchical deep reinforcement learning architecture that integrates intrinsic motivation and temporal abstraction for better exploration.

Findings

Effective in sparse reward environments like Montezuma's Revenge

Outperforms baseline methods in complex decision processes

Demonstrates flexible goal specification capabilities

Abstract

Learning goal-directed behavior in environments with sparse feedback is a major challenge for reinforcement learning algorithms. The primary difficulty arises due to insufficient exploration, resulting in an agent being unable to learn robust value functions. Intrinsically motivated agents can explore new behavior for its own sake rather than to directly solve problems. Such intrinsic behaviors could eventually help the agent solve tasks posed by the environment. We present hierarchical-DQN (h-DQN), a framework to integrate hierarchical value functions, operating at different temporal scales, with intrinsically motivated deep reinforcement learning. A top-level value function learns a policy over intrinsic goals, and a lower-level function learns a policy over atomic actions to satisfy the given goals. h-DQN allows for flexible goal specifications, such as functions over entities and relations. This provides an efficient space for exploration in complicated environments. We demonstrate the strength of our approach on two problems with very sparse, delayed feedback: (1) a complex discrete stochastic decision process, and (2) the classic ATARI game `Montezuma's Revenge'.