Misclassification excess risk bounds for PAC-Bayesian classification via convexified loss

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autore principale: Mai, The Tien
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909289401024512
author Mai, The Tien
author_facet Mai, The Tien
contents PAC-Bayesian bounds have proven to be a valuable tool for deriving generalization bounds and for designing new learning algorithms in machine learning. However, it typically focus on providing generalization bounds with respect to a chosen loss function. In classification tasks, due to the non-convex nature of the 0-1 loss, a convex surrogate loss is often used, and thus current PAC-Bayesian bounds are primarily specified for this convex surrogate. This work shifts its focus to providing misclassification excess risk bounds for PAC-Bayesian classification when using a convex surrogate loss. Our key ingredient here is to leverage PAC-Bayesian relative bounds in expectation rather than relying on PAC-Bayesian bounds in probability. We demonstrate our approach in several important applications.
format Preprint
id arxiv_https___arxiv_org_abs_2408_08675
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Misclassification excess risk bounds for PAC-Bayesian classification via convexified loss
Mai, The Tien
Machine Learning
PAC-Bayesian bounds have proven to be a valuable tool for deriving generalization bounds and for designing new learning algorithms in machine learning. However, it typically focus on providing generalization bounds with respect to a chosen loss function. In classification tasks, due to the non-convex nature of the 0-1 loss, a convex surrogate loss is often used, and thus current PAC-Bayesian bounds are primarily specified for this convex surrogate. This work shifts its focus to providing misclassification excess risk bounds for PAC-Bayesian classification when using a convex surrogate loss. Our key ingredient here is to leverage PAC-Bayesian relative bounds in expectation rather than relying on PAC-Bayesian bounds in probability. We demonstrate our approach in several important applications.
title Misclassification excess risk bounds for PAC-Bayesian classification via convexified loss
topic Machine Learning
url https://arxiv.org/abs/2408.08675