End-to-End Training of Deep Visuomotor Policies

This seems like the first end-to-end imitation learning strategy? really old paper.