Fetching the paper…

MaDi: Learning to Mask Distractions for Generalization in Visual Deep Reinforcement Learning · Around