Bayesian Coresets and Edward: parallel computation

less than 1 minute read


In a previous posts we have explored the use of Bayesian coreset on synthetic data and its application to a phishing data set. We replicated experiments from the original article by integrating the original code with the Edward framework.

In this notebook, we continue our study of the original article by evaluating the computation of Bayesian coresets in parallel. We use again a synthetic data set to explore the results produced by Bayesian coresets when running on identically-distributed sub-sets and differently-distributed sub-sets.

Notebook here