Stochastic Variational BayesΒΆ
SVB
is a package to perform stochastic Bayesian inference on a nonlinear
forward model (i.e. a parameterised model which is able to predict data
values from a set of parameter values).
The implementation leverages the TensorFlow framework to perform efficient optimisation of the model parameters given an experimental data set.