Three rates of convergence or separation via U-statistics in a dependent framework

TL;DR

This paper leverages a recent concentration inequality for U-statistics in dependent Markov chains to advance spectral estimation, online learning, and goodness-of-fit testing in dependent data settings.

Contribution

It introduces a new exponential inequality for spectral estimation with kernels of mixed eigenvalues, and applies it to online learning and goodness-of-fit testing in Markov chain frameworks.

Findings

New exponential inequality for spectral estimation with mixed eigenvalues

Online-to-batch conversion for Markov chain data

Non-asymptotic goodness-of-fit test with prescribed power

Abstract

Despite the ubiquity of U-statistics in modern Probability and Statistics, their non-asymptotic analysis in a dependent framework may have been overlooked. In a recent work, a new concentration inequality for U-statistics of order two for uniformly ergodic Markov chains has been proved. In this paper, we put this theoretical breakthrough into action by pushing further the current state of knowledge in three different active fields of research. First, we establish a new exponential inequality for the estimation of spectra of trace class integral operators with MCMC methods. The novelty is that this result holds for kernels with positive and negative eigenvalues, which is new as far as we know. In addition, we investigate generalization performance of online algorithms working with pairwise loss functions and Markov chain samples. We provide an online-to-batch conversion result by showing how we can extract a low risk hypothesis from the sequence of hypotheses generated by any online learner. We finally give a non-asymptotic analysis of a goodness-of-fit test on the density of the invariant measure of a Markov chain. We identify some classes of alternatives over which our test based on the $L_2$ distance has a prescribed power.