contribution is to introduce neural ODEs in a model-based RL framework for SMDPs: our experiments help us