Continuous-time on-policy neural Reinforcement Learning of working memory tasks | IEEE Conference Publication | IEEE Xplore