On the Expressiveness of Approximate Inference in Bayesian Neural Networks

Open in new window