Debiased Sinkhorn barycenters

Control of autonomous ordinary differential equations In short, we have a method using piecewise uniform cubic B´ ezier curves to approximate optimal trajectories.. This method can

This quantity is usually estimated with the plug-in estimator, defined via a discrete optimal transport problem which can be solved to ε-accuracy by adding an entropic regularization

We emphasize that this paper introduces two different approaches (Sliced and Radon) together with their corre- sponding discretization (Lagrangian and Eulerian) to cope with the

We consider the transfer operators of non-uniformly expanding maps for potentials of various regularity, and show that a specific property of potentials (“flatness”) implies

In all these cases, our result provides the existence of the target mean distribution while the consistency results allow the barycenters to be approximated by taking the barycenter

To define probability models where averaging in the Wasserstein space amounts to take the expectation of an optimal transport map, we study statistical models for which the notion

Convergence of an estimator of the Wasserstein distance between two continuous probability distributions.. Thierry Klein, Jean-Claude Fort,

In this work we characterize the differential properties of the original Sinkhorn distance, proving that it enjoys the same smoothness as its regularized version and we