Using Contrastive Samples for Identifying and Leveraging Possible Causal Relationships in Reinforcement Learning

TL;DR

This paper introduces Contrastive Experience Replay (CER), a method that identifies potential causal transitions in reinforcement learning by analyzing deviations in states and rewards, and enhances learning by incorporating contrastive samples.

Contribution

The paper proposes CER, a novel approach that leverages causal inference concepts in reinforcement learning to improve performance on navigation tasks.

Findings

CER outperforms standard value-based methods on 2D navigation tasks.

Including contrastive samples enhances the learning process.

CER is applicable to various off-policy reinforcement learning algorithms.

Abstract

A significant challenge in reinforcement learning is quantifying the complex relationship between actions and long-term rewards. The effects may manifest themselves over a long sequence of state-action pairs, making them hard to pinpoint. In this paper, we propose a method to link transitions with significant deviations in state with unusually large variations in subsequent rewards. Such transitions are marked as possible causal effects, and the corresponding state-action pairs are added to a separate replay buffer. In addition, we include \textit{contrastive} samples corresponding to transitions from a similar state but with differing actions. Including this Contrastive Experience Replay (CER) during training is shown to outperform standard value-based methods on 2D navigation tasks. We believe that CER can be useful for a broad class of learning tasks, including for any off-policy reinforcement learning algorithm.