On the optimization and pruning for Bayesian deep learning

Xiongwen Ke, Yanan Fan

The goal of Bayesian deep learning is to provide uncertainty quantification via the posterior distribution. However, exact inference over the weight space is computationally intrac…