Kernelized Wasserstein Natural Gradient
Skip to main content
eScholarship
Open Access Publications from the University of California

Kernelized Wasserstein Natural Gradient

  • Author(s): Arbel, Michael
  • Gretton, Arthur
  • Li, Wuchen
  • Montufar, Guido
  • et al.

Published Web Location

https://openreview.net/pdf?id=Hklz71rYvS
No data is associated with this publication.
Abstract

Many machine learning problems can be expressed as the optimization of some cost functional over a parametric family of probability distributions. It is often beneficial to solve such optimization problems using natural gradient methods. These methods are invariant to the parametrization of the family, and thus can yield more effective optimization. Unfortunately, computing the natural gradient is challenging as it requires inverting a high dimensional matrix at each iteration. We propose a general framework to approximate the natural gradient for the Wasserstein metric, by leveraging a dual formulation of the metric restricted to a Reproducing Kernel Hilbert Space. Our approach leads to an estimator for gradient direction that can trade-off accuracy and computational cost, with theoretical guarantees. We verify its accuracy on simple examples, and show the advantage of using such an estimator in classification tasks on Cifar10 and Cifar100 empirically.

Many UC-authored scholarly publications are freely available on this site because of the UC's open access policies. Let us know how this access is important for you.

Item not freely available? Link broken?
Report a problem accessing this item