Boolean Decision Rules for Reinforcement Learning Policy Summarisation

TL;DR

This paper introduces a Boolean decision rule-based method for post-hoc summarization of reinforcement learning policies to enhance explainability and safety, demonstrated on a gridworld environment.

Contribution

It proposes a novel rule-based approach for summarizing RL policies, aiding interpretability and safety considerations in reinforcement learning.

Findings

Simple rules effectively summarize RL policies

Rules can be used to impose safety constraints

Method applied successfully to a gridworld environment

Abstract

Explainability of Reinforcement Learning (RL) policies remains a challenging research problem, particularly when considering RL in a safety context. Understanding the decisions and intentions of an RL policy offer avenues to incorporate safety into the policy by limiting undesirable actions. We propose the use of a Boolean Decision Rules model to create a post-hoc rule-based summary of an agent's policy. We evaluate our proposed approach using a DQN agent trained on an implementation of a lava gridworld and show that, given a hand-crafted feature representation of this gridworld, simple generalised rules can be created, giving a post-hoc explainable summary of the agent's policy. We discuss possible avenues to introduce safety into a RL agent's policy by using rules generated by this rule-based model as constraints imposed on the agent's policy, as well as discuss how creating simple rule summaries of an agent's policy may help in the debugging process of RL agents.