Robust Visual Domain Randomization for Reinforcement Learning.

By Reda Bahi Slaoui, William R. Clements, Jakob N. Foerster, S'ebastien Toth. 2019

20 citations recorded by the source corpus.

Producing agents that can generalize to a wide range of visually different environments is a significant challenge in reinforcement learning. One method for overcoming this issue is visual domain randomization, whereby at the start of each training episode some visual aspects of the environment are randomized so that the agent is exposed to many possible variations. However, domain randomization is highly inefficient and may lead to policies with high variance across domains. Instead, we propose a regularization method whereby the agent is only trained on one variation of the environment, and its learned state representations are regularized during training to be invariant across domains. We conduct experiments that demonstrate that our technique leads to more efficient and robust learning than standard domain randomization, while achieving equal generalization scores.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

Citations in the corpus

• Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixels (2020, neutral): “There are also domain randomization techniques [52, 49] which synthetically apply similar variations, but assume control of the data generation procedure, in contrast to our method.”