Cross-domain Imitation from Observations

       Dripta S. Raychaudhuri * 1 Sujoy Paul * 2  Jeroen van Baar 3 Amit K. Roy-Chowdhury 1

                                    Expert domain         Proxy task       Inference task
             Abstract
   Imitation learning seeks to circumvent the diffi-
   culty in designing proper reward functions for
   training agents by utilizing expert behavior. With
   environments modeled as Markov Decision Pro-                                      Transform
...