I've noticed that DeepMind completely ignores imitation learning area of research. I never saw them citing any of the fundamental papers, despite a bunch of their reinforcement learning ideas explicitly match those ideas that are now 8+ years old.