Deep Reinforcement Learning by Parallelizing Reward and Punishment using the MaxPain Architecture | IEEE Conference Publication | IEEE Xplore