On First-Order Meta-Reinforcement Learning with Moreau Envelopes | IEEE Conference Publication | IEEE Xplore