Safe Reinforcement Learning via Probabilistic Timed Computation Tree Logic | IEEE Conference Publication | IEEE Xplore