Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixels

TL;DR

This paper introduces a simple data augmentation technique that enhances model-free reinforcement learning from pixel inputs, achieving state-of-the-art results without auxiliary losses or pre-training.

Contribution

The proposed augmentation method significantly improves the performance of standard model-free RL algorithms like SAC on pixel-based tasks, surpassing existing model-based and contrastive learning approaches.

Findings

Dramatic performance improvement of SAC with augmentation

State-of-the-art results on DeepMind control suite

Compatible with any model-free RL algorithm

Abstract

We propose a simple data augmentation technique that can be applied to standard model-free reinforcement learning algorithms, enabling robust learning directly from pixels without the need for auxiliary losses or pre-training. The approach leverages input perturbations commonly used in computer vision tasks to regularize the value function. Existing model-free approaches, such as Soft Actor-Critic (SAC), are not able to train deep networks effectively from image pixels. However, the addition of our augmentation method dramatically improves SAC's performance, enabling it to reach state-of-the-art performance on the DeepMind control suite, surpassing model-based (Dreamer, PlaNet, and SLAC) methods and recently proposed contrastive learning (CURL). Our approach can be combined with any model-free reinforcement learning algorithm, requiring only minor modifications. An implementation can be found at https://sites.google.com/view/data-regularized-q.