Tutorial on Bayesian Neural ODE: SGLD Sampler
In this tutorial, we show how the DiffEqFlux.jl library in Julia can be seamlessly combined with Bayesian estimation libraries like AdvancedHMC.jl and Turing.jl. This enables converting Neural ODEs to Bayesian Neural ODEs, which enables us to estimate the error in the Neural ODE estimation and forecasting. In this tutorial, a working example of the Bayesian Neural ODE: SGLD sampler is shown.