Generating Explanations from Deep Reinforcement Learning Using Episodic Memory

TL;DR

This paper introduces a method for Deep Reinforcement Learning agents to generate human-readable explanations by using episodic memory to identify key decisions, thereby improving interpretability and accelerating learning.

Contribution

The paper proposes integrating episodic memory into Deep RL to produce concise explanations and enhance learning efficiency, a novel approach for interpretability in RL.

Findings

Episodic memory helps identify key decisions for explanations.

Generated explanations are human-readable and concise.

Method accelerates learning in naive Deep RL agents.

Abstract

Deep Reinforcement Learning (RL) involves the use of Deep Neural Networks (DNNs) to make sequential decisions in order to maximize reward. For many tasks the resulting sequence of actions produced by a Deep RL policy can be long and difficult to understand for humans. A crucial component of human explanations is selectivity, whereby only key decisions and causes are recounted. Imbuing Deep RL agents with such an ability would make their resulting policies easier to understand from a human perspective and generate a concise set of instructions to aid the learning of future agents. To this end we use a Deep RL agent with an episodic memory system to identify and recount key decisions during policy execution. We show that these decisions form a short, human readable explanation that can also be used to speed up the learning of naive Deep RL agents in an algorithm-independent manner.