Exponential Concentration for Geometric-Median-of-Means in Non-Positive Curvature Spaces

TL;DR

This paper extends median-of-means estimators to non-positive curvature metric spaces, achieving exponential concentration for the population Fréchet mean under minimal moment conditions, surpassing the polynomial concentration of empirical means.

Contribution

It introduces new median-of-means notions and inequalities in general metric spaces, providing the first non-asymptotic concentration results in non-vector, non-compact, infinite-dimensional spaces.

Findings

Median-of-means estimators achieve exponential concentration.

Empirical Fréchet mean has polynomial concentration.

Results apply to spaces with non-positive Alexandrov curvature.

Abstract

In Euclidean spaces, the empirical mean vector as an estimator of the population mean is known to have polynomial concentration unless a strong tail assumption is imposed on the underlying probability measure. The idea of median-of-means tournament has been considered as a way of overcoming the sub-optimality of the empirical mean vector. In this paper, to address the sub-optimal performance of the empirical mean in a more general setting, we consider general Polish spaces with a general metric, which are allowed to be non-compact and of infinite-dimension. We discuss the estimation of the associated population Frechet mean, and for this we extend the existing notion of median-of-means to this general setting. We devise several new notions and inequalities associated with the geometry of the underlying metric, and using them we study the concentration properties of the extended notions of median-of-means as the estimators of the population Frechet mean. We show that the new estimators achieve exponential concentration under only a second moment condition on the underlying distribution, while the empirical Frechet mean has polynomial concentration. We focus our study on spaces with non-positive Alexandrov curvature since they afford slower rates of convergence than spaces with positive curvature. We note that this is the first work that derives non-asymptotic concentration inequalities for extended notions of the median-of-means in non-vector spaces with a general metric.