The supplement includes the following files:

highway.avi - video showing policies learned by GPIRL, MaxEnt, and FIRL on the highway environment from human demonstrations of the "police" policy; the first clip shows one of eight human demonstrations, and the second clip shows the corresponding optimal policy derived from the true reward function

gpirl_supplement.pdf - appendices that provide a derivation of the partial derivatives of each component of the GPIRL log likelihood, as well as details about the warped kernel function and our random restarts procedure

gpirl_result_tables.pdf - tables containing the complete results for all algorithms that were tested on each experiment described in the paper

csv/*.csv - tables containing the complete results, provided as csv files for convenience
