ELBO, regularized maximum likelihood, and their common one-sample approximation for training stochastic neural networks.

Sina Dubener, Simon Damm, Asja Fischer

Browse the full UAI paper archive.