Reinforcement Learning with Augmented Data

TL;DR

RAD enhances reinforcement learning by applying data augmentation techniques, significantly improving data-efficiency and generalization across various benchmarks for pixel-based and state-based inputs.

Contribution

Introduces RAD, a versatile data augmentation module for RL, and systematically studies its impact, establishing new state-of-the-art results.

Findings

RAD outperforms existing methods on DeepMind Control Suite.

RAD improves test-time generalization on ProcGen benchmarks.

Simple augmentations can surpass complex RL algorithms.

Abstract

Learning from visual observations is a fundamental yet challenging problem in Reinforcement Learning (RL). Although algorithmic advances combined with convolutional neural networks have proved to be a recipe for success, current methods are still lacking on two fronts: (a) data-efficiency of learning and (b) generalization to new environments. To this end, we present Reinforcement Learning with Augmented Data (RAD), a simple plug-and-play module that can enhance most RL algorithms. We perform the first extensive study of general data augmentations for RL on both pixel-based and state-based inputs, and introduce two new data augmentations - random translate and random amplitude scale. We show that augmentations such as random translate, crop, color jitter, patch cutout, random convolutions, and amplitude scale can enable simple RL algorithms to outperform complex state-of-the-art methods across common benchmarks. RAD sets a new state-of-the-art in terms of data-efficiency and final performance on the DeepMind Control Suite benchmark for pixel-based control as well as OpenAI Gym benchmark for state-based control. We further demonstrate that RAD significantly improves test-time generalization over existing methods on several OpenAI ProcGen benchmarks. Our RAD module and training code are available at https://www.github.com/MishaLaskin/rad.