On the Benefit of Optimal Transport for Curriculum Reinforcement Learning | IEEE Journals & Magazine | IEEE Xplore