Doubly Reparameterized Gradient Estimators for Monte Carlo Objectives
international conference on learning representations, 2019.
Deep latent variable models have become a popular model choice due to the scalable learning algorithms introduced by (Kingma u0026 Welling, 2013; Rezende et al., 2014). These approaches maximize a variational lower bound on the intractable log likelihood of the observed data. Burda et al. (2015) introduced a multi-sample variational bound...More
PPT (Upload PPT)