Using Reinforcement Learning for Variational Inference

How to use RL to make inferences in sequential settings.