Model-Free Reinforcement Learning for Spatiotemporal Tasks Using Symbolic Automata | IEEE Conference Publication | IEEE Xplore