Automatic Data Augmentation for Generalization in Deep RL
We compare three approaches for automatically finding an appropriate augmentation combined with two novel regularization terms for the policy and value ...
data-augmentation reinforcement-learning kornia pytorch research paper code arxiv:2006.12862

Deep reinforcement learning (RL) agents often fail to generalize to unseen scenarios, even when they are trained on many instances of semantically similar environments. Data augmentation has recently been shown to improve the sample efficiency and generalization of RL agents. However, different tasks tend to benefit from different kinds of data augmentation. In this paper, we compare three approaches for automatically finding an appropriate augmentation. These are combined with two novel regularization terms for the policy and value function, required to make the use of data augmentation theoretically sound for certain actor-critic algorithms. We evaluate our methods on the Procgen benchmark which consists of 16 procedurally-generated environments and show that it improves test performance by ~40% relative to standard RL algorithms. Our agent outperforms other baselines specifically designed to improve generalization in RL. In addition, we show that our agent learns policies and representations that are more robust to changes in the environment that do not affect the agent, such as the background.

Don't forget to tag @rraileanu in your comment, otherwise they may not be notified.

Authors community post
Share this project
Similar projects
Data augmentation recipes in tf.keras image-based models
Learn about different ways of doing data augmentation when training an image classifier in tf.keras.
Semi-Supervised Learning in Computer Vision
A comprehensive overview of recent semi-supervised learning methods in Computer Vision
Classify photos in 600 classes using nine million Open Images
If you’re looking build an image classifier but need training data, look no further than Google Open Images.
DeltaPy⁠⁠
Tabular Data Augmentation & Feature Engineering.