Concentration in unbounded metric spaces and algorithmic stability
Probability
2013-09-12 v2 Machine Learning
Functional Analysis
Abstract
We prove an extension of McDiarmid's inequality for metric spaces with unbounded diameter. To this end, we introduce the notion of the {\em subgaussian diameter}, which is a distribution-dependent refinement of the metric diameter. Our technique provides an alternative approach to that of Kutin and Niyogi's method of weakly difference-bounded functions, and yields nontrivial, dimension-free results in some interesting cases where the former does not. As an application, we give apparently the first generalization bound in the algorithmic stability setting that holds for unbounded loss functions. We furthermore extend our concentration inequality to strongly mixing processes.
Cite
@article{arxiv.1309.1007,
title = {Concentration in unbounded metric spaces and algorithmic stability},
author = {Aryeh Kontorovich},
journal= {arXiv preprint arXiv:1309.1007},
year = {2013}
}