Self-Supervised Imitation Learning

Pierre Sermanet¹*, Corey Lynch¹*†, Yevgen Chebotar²*, Kelvin Xu¹†, Jasmine Hsu¹, Eric Jang¹, Stefan Schaal², Sergey Levine^1
1 Google Brain, ² University of Southern California
(* equal contribution, † Google Brain Residency program g.co/brainresidency)

Project

The general goal of this project is to give to robots the ability to learn to imitate humans from observation and without labels.
We present the following papers as steps toward that goal.

1. Time-Contrastive Networks (TCN)

2. Unsupervised Perceptual Rewards

3. TCN update: Learning to imitate