A Deep Reinforcement Learning Heuristic for SAT based on Antagonist Graph Neural Networks | IEEE Conference Publication | IEEE Xplore