ResearchBenchmarksReinforcement Learning on OpenAI Gym AntReinforcement Learning on OpenAI Gym AntFollow73,500Number of EpisodesTRPO+NN28,26040,00551,75063,495Mar 8, 2017Evaluation resultsMethodMethodLinksNumber of EpisodesTRPO+NNThreshold (Th.)=3580Threshold (Th.)=35802017.0373,500LinearThreshold (Th.)=3580Threshold (Th.)=35802017.0339,240RBFThreshold (Th.)=3580Threshold (Th.)=35802017.0330,000