Scalable Semantic Non-Markovian Simulation Proxy for Reinforcement Learning | IEEE Conference Publication | IEEE Xplore