Using Part-based Representations for Explainable Deep Reinforcement Learning

TL;DR

This paper introduces a non-negative training method for deep reinforcement learning that facilitates interpretable part-based representations, demonstrated on the Cartpole benchmark.

Contribution

It proposes a novel non-negative initialization and sign-preserving training approach to improve interpretability in RL models with part-based representations.

Findings

Enhanced interpretability of RL models through part-based representations.

Improved training stability and convergence with the proposed method.

Successful application on the Cartpole benchmark.

Abstract

Utilizing deep learning models to learn part-based representations holds significant potential for interpretable-by-design approaches, as these models incorporate latent causes obtained from feature representations through simple addition. However, training a part-based learning model presents challenges, particularly in enforcing non-negative constraints on the model's parameters, which can result in training difficulties such as instability and convergence issues. Moreover, applying such approaches in Deep Reinforcement Learning (RL) is even more demanding due to the inherent instabilities that impact many optimization methods. In this paper, we propose a non-negative training approach for actor models in RL, enabling the extraction of part-based representations that enhance interpretability while adhering to non-negative constraints. To this end, we employ a non-negative initialization technique, as well as a modified sign-preserving training method, which can ensure better gradient flow compared to existing approaches. We demonstrate the effectiveness of the proposed approach using the well-known Cartpole benchmark.