March 1, 2021 AI learns to speedrun QWOP — combining reinforcement and imitation learning to get a top 10 speedrun (via) #
