System Design Perspective for Human-Level Agents Using Deep Reinforcement Learning: A Survey | IEEE Journals & Magazine | IEEE Xplore