18
2

Hyperbolic contractivity and the Hilbert metric on probability measures

Abstract

This paper gives a self-contained introduction to the Hilbert projective metric H\mathcal{H} and its fundamental properties, with a particular focus on the space of probability measures. We start by defining the Hilbert pseudo-metric on convex cones, focusing mainly on dual formulations of H\mathcal{H} . We show that linear operators on convex cones contract in the distance given by the hyperbolic tangent of H\mathcal{H}, which in particular implies Birkhoff's classical contraction result for H\mathcal{H}. Turning to spaces of probability measures, where H\mathcal{H} is a metric, we analyse the dual formulation of H\mathcal{H} in the general setting, and explore the geometry of the probability simplex under H\mathcal{H} in the special case of discrete probability measures. Throughout, we compare H\mathcal{H} with other distances between probability measures. In particular, we show how convergence in H\mathcal{H} implies convergence in total variation, pp-Wasserstein distance, and any ff-divergence. Furthermore, we derive a novel sharp bound for the total variation between two probability measures in terms of their Hilbert distance.

View on arXiv
Comments on this paper