Action Candidate Driven Clipped Double Q-Learning for Discrete and Continuous Action Tasks | IEEE Journals & Magazine | IEEE Xplore