Sequential Action-Induced Invariant Representation for Reinforcement Learning

TL;DR

This paper introduces SAR, a method that leverages sequential action information to learn robust, distraction-invariant representations in visual reinforcement learning, improving performance in complex, noisy environments.

Contribution

The paper proposes a novel SAR approach that incorporates sequential actions into representation learning to enhance task relevance and robustness against distractions.

Findings

Achieves state-of-the-art performance on DeepMind Control with distractions.

Effectively disregards task-irrelevant information in autonomous driving.

Demonstrates improved generalization and robustness in noisy environments.

Abstract

How to accurately learn task-relevant state representations from high-dimensional observations with visual distractions is a realistic and challenging problem in visual reinforcement learning. Recently, unsupervised representation learning methods based on bisimulation metrics, contrast, prediction, and reconstruction have shown the ability for task-relevant information extraction. However, due to the lack of appropriate mechanisms for the extraction of task information in the prediction, contrast, and reconstruction-related approaches and the limitations of bisimulation-related methods in domains with sparse rewards, it is still difficult for these methods to be effectively extended to environments with distractions. To alleviate these problems, in the paper, the action sequences, which contain task-intensive signals, are incorporated into representation learning. Specifically, we propose a Sequential Action--induced invariant Representation (SAR) method, in which the encoder is optimized by an auxiliary learner to only preserve the components that follow the control signals of sequential actions, so the agent can be induced to learn the robust representation against distractions. We conduct extensive experiments on the DeepMind Control suite tasks with distractions while achieving the best performance over strong baselines. We also demonstrate the effectiveness of our method at disregarding task-irrelevant information by deploying SAR to real-world CARLA-based autonomous driving with natural distractions. Finally, we provide the analysis results of generalization drawn from the generalization decay and t-SNE visualization. Code and demo videos are available at https://github.com/DMU-XMU/SAR.git.