Poster
Sticking the Landing: Simple, Lower-Variance Gradient Estimators for Variational Inference
Geoffrey Roeder · Yuhuai Wu · David Duvenaud
Pacific Ballroom #180
Keywords: [ Probabilistic Methods ] [ Deep Autoencoders ] [ Variational Inference ] [ Stochastic Methods ]
We propose a simple and general variant of the standard reparameterized gradient estimator for the variational evidence lower bound. Specifically, we remove a part of the total derivative with respect to the variational parameters that corresponds to the score function. Removing this term produces an unbiased gradient estimator whose variance approaches zero as the approximate posterior approaches the exact posterior. We analyze the behavior of this gradient estimator theoretically and empirically, and generalize it to more complex variational distributions such as mixtures and importance-weighted posteriors.
Live content is unavailable. Log in and register to view live content