Saved in:
Bibliographic Details
Main Authors: Abeles, Baptiste, Clerico, Eugenio, Neu, Gergely
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2406.12600
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929719735222272
author Abeles, Baptiste
Clerico, Eugenio
Neu, Gergely
author_facet Abeles, Baptiste
Clerico, Eugenio
Neu, Gergely
contents We study the generalization error of statistical learning algorithms in a non-i.i.d. setting, where the training data is sampled from a stationary mixing process. We develop an analytic framework for this scenario based on a reduction to online learning with delayed feedback. In particular, we show that the existence of an online learning algorithm with bounded regret (against a fixed statistical learning algorithm in a specially constructed game of online learning with delayed feedback) implies low generalization error of said statistical learning method even if the data sequence is sampled from a mixing time series. The rates demonstrate a trade-off between the amount of delay in the online learning game and the degree of dependence between consecutive data points, with near-optimal rates recovered in a number of well-studied settings when the delay is tuned appropriately as a function of the mixing time of the process.
format Preprint
id arxiv_https___arxiv_org_abs_2406_12600
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Generalization bounds for mixing processes via delayed online-to-PAC conversions
Abeles, Baptiste
Clerico, Eugenio
Neu, Gergely
Machine Learning
We study the generalization error of statistical learning algorithms in a non-i.i.d. setting, where the training data is sampled from a stationary mixing process. We develop an analytic framework for this scenario based on a reduction to online learning with delayed feedback. In particular, we show that the existence of an online learning algorithm with bounded regret (against a fixed statistical learning algorithm in a specially constructed game of online learning with delayed feedback) implies low generalization error of said statistical learning method even if the data sequence is sampled from a mixing time series. The rates demonstrate a trade-off between the amount of delay in the online learning game and the degree of dependence between consecutive data points, with near-optimal rates recovered in a number of well-studied settings when the delay is tuned appropriately as a function of the mixing time of the process.
title Generalization bounds for mixing processes via delayed online-to-PAC conversions
topic Machine Learning
url https://arxiv.org/abs/2406.12600