A Bayesian Interpretation of Recurrence in Neural Networks - View it on GitHub