Variation Due to Regularization Tractably Recovers Bayesian Deep Learning Uncertainty.

James McInerney, Nathan Kallus

Browse the full AISTATS paper archive.