Skip to main content

Reducing reparameterization gradient variance

Author(s): Miller, AC; Foti, NJ; D Amour, A; Adams, Ryan P

To refer to this page use:
Abstract: Optimization with noisy gradients has become ubiquitous in statistics and machine learning. Reparameterization gradients, or gradient estimates computed via the "reparameterization trick," represent a class of noisy gradients often used in Monte Carlo variational inference (MCVI). However, when these gradient estimators are too noisy, the optimization procedure can be slow or fail to converge. One way to reduce noise is to generate more samples for the gradient estimate, but this can be computationally expensive. Instead, we view the noisy gradient as a random variable, and form an inexpensive approximation of the generating procedure for the gradient sample. This approximation has high correlation with the noisy gradient by construction, making it a useful control variate for variance reduction. We demonstrate our approach on a non-conjugate hierarchical model and a Bayesian neural net where our method attained orders of magnitude (20-2, 000×) reduction in gradient variance resulting in faster and more stable optimization.
Publication Date: 2017
Citation: Miller, AC, Foti, NJ, D Amour, A, Adams, RP. (2017). Reducing reparameterization gradient variance. 2017-December (3709 - 3719
Pages: 3709 - 3719
Type of Material: Conference Article
Journal/Proceeding Title: Advances in Neural Information Processing Systems
Version: Author's manuscript

Items in OAR@Princeton are protected by copyright, with all rights reserved, unless otherwise indicated.