Solving Continuous Control with Episodic Memory

TL;DR

This paper introduces a novel episodic memory-based reinforcement learning algorithm for continuous control tasks, enhancing sample efficiency by integrating episodic memory with Actor-Critic methods and prioritizing experience replay.

Contribution

It presents a new approach combining episodic memory with Actor-Critic architecture and episodic-based replay prioritization for continuous control, demonstrating improved sample efficiency.

Findings

Achieves greater sample-efficiency than state-of-the-art algorithms

Successfully applies episodic memory to continuous control tasks

Shows improved performance on OpenAI gym benchmarks

Abstract

Episodic memory lets reinforcement learning algorithms remember and exploit promising experience from the past to improve agent performance. Previous works on memory mechanisms show benefits of using episodic-based data structures for discrete action problems in terms of sample-efficiency. The application of episodic memory for continuous control with a large action space is not trivial. Our study aims to answer the question: can episodic memory be used to improve agent's performance in continuous control? Our proposed algorithm combines episodic memory with Actor-Critic architecture by modifying critic's objective. We further improve performance by introducing episodic-based replay buffer prioritization. We evaluate our algorithm on OpenAI gym domains and show greater sample-efficiency compared with the state-of-the art model-free off-policy algorithms.