Continuous Episodic Control

TL;DR

Continuous Episodic Control (CEC) is a new non-parametric method for reinforcement learning in continuous action spaces, enabling faster learning in sparse-reward environments while maintaining strong long-term performance.

Contribution

This paper introduces CEC, the first non-parametric episodic memory algorithm for continuous action spaces, extending episodic control to a broader class of problems.

Findings

CEC learns faster than state-of-the-art algorithms

CEC maintains good long-term performance

CEC is effective in sparse-reward continuous control environments

Abstract

Non-parametric episodic memory can be used to quickly latch onto high-rewarded experience in reinforcement learning tasks. In contrast to parametric deep reinforcement learning approaches in which reward signals need to be back-propagated slowly, these methods only need to discover the solution once, and may then repeatedly solve the task. However, episodic control solutions are stored in discrete tables, and this approach has so far only been applied to discrete action space problems. Therefore, this paper introduces Continuous Episodic Control (CEC), a novel non-parametric episodic memory algorithm for sequential decision making in problems with a continuous action space. Results on several sparse-reward continuous control environments show that our proposed method learns faster than state-of-the-art model-free RL and memory-augmented RL algorithms, while maintaining good long-run performance as well. In short, CEC can be a fast approach for learning in continuous control tasks.