Adapting control policies from simulation to reality using a pairwise loss

Open in new window