The application of reinforcement learning (RL) algorithms is often hindered by the combinatorial explosion of the state space. Previous works have leveraged abstractions which condense large state spaces to find tractable ...