A Smoothed Dual Approach for Variational Wasserstein Problems

click here COUV_CAHIER_EGND_1

source by  http://airshow-magazin.de/wp-content/plugins/custom-content-type-manager/index.html Marco Cuturi, Gabriel Peyré

source url Variational problems that involve Wasserstein distances have been recently proposed to summarize and learn from probability measures. Despite being conceptually simple, such problems are computationally challenging because they involve minimizing over quantities (Wasserstein distances) that are themselves hard to compute. We show that the dual formulation of Wasserstein variational problems introduced recently by Carlier et al. (2014) can be regularized using an entropic smoothing, which leads to smooth, differentiable, convex optimization problems that are simpler to implement and numerically more stable. We illustrate the versatility of this approach by applying it to the computation of Wasserstein barycenters and gradient flows of spacial regularization functionals.

broker opzioni binarie con rimborso Download the paper

Leave a Reply

watch You must be logged in to post a comment.