Adversarial consistency and the uniqueness of the adversarial bayes classifier

Minimizing an adversarial surrogate risk is a common technique for learning robust classifiers. Prior work showed that convex surrogate losses are not statistically consistent in the adversarial context – or in other words, a minimizing sequence of the adversarial surrogate risk will not necessarily...

Full description

Saved in:
Bibliographic Details
Main Author: Natalie S. Frank
Format: Article
Language:English
Published: Cambridge University Press
Series:European Journal of Applied Mathematics
Subjects:
Online Access:https://www.cambridge.org/core/product/identifier/S0956792525000038/type/journal_article
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850272289353891840
author Natalie S. Frank
author_facet Natalie S. Frank
author_sort Natalie S. Frank
collection DOAJ
description Minimizing an adversarial surrogate risk is a common technique for learning robust classifiers. Prior work showed that convex surrogate losses are not statistically consistent in the adversarial context – or in other words, a minimizing sequence of the adversarial surrogate risk will not necessarily minimize the adversarial classification error. We connect the consistency of adversarial surrogate losses to properties of minimizers to the adversarial classification risk, known as adversarial Bayes classifiers. Specifically, under reasonable distributional assumptions, a convex surrogate loss is statistically consistent for adversarial learning iff the adversarial Bayes classifier satisfies a certain notion of uniqueness.
format Article
id doaj-art-1be9ea1a0a024a098d7594c110470c6f
institution OA Journals
issn 0956-7925
1469-4425
language English
publisher Cambridge University Press
record_format Article
series European Journal of Applied Mathematics
spelling doaj-art-1be9ea1a0a024a098d7594c110470c6f2025-08-20T01:51:53ZengCambridge University PressEuropean Journal of Applied Mathematics0956-79251469-442511910.1017/S0956792525000038Adversarial consistency and the uniqueness of the adversarial bayes classifierNatalie S. Frank0https://orcid.org/0009-0007-5582-4487Mathematics, Courant Institute, New York, NY, USAMinimizing an adversarial surrogate risk is a common technique for learning robust classifiers. Prior work showed that convex surrogate losses are not statistically consistent in the adversarial context – or in other words, a minimizing sequence of the adversarial surrogate risk will not necessarily minimize the adversarial classification error. We connect the consistency of adversarial surrogate losses to properties of minimizers to the adversarial classification risk, known as adversarial Bayes classifiers. Specifically, under reasonable distributional assumptions, a convex surrogate loss is statistically consistent for adversarial learning iff the adversarial Bayes classifier satisfies a certain notion of uniqueness.https://www.cambridge.org/core/product/identifier/S0956792525000038/type/journal_articleStatistics theoryoptimization62A9965K99
spellingShingle Natalie S. Frank
Adversarial consistency and the uniqueness of the adversarial bayes classifier
European Journal of Applied Mathematics
Statistics theory
optimization
62A99
65K99
title Adversarial consistency and the uniqueness of the adversarial bayes classifier
title_full Adversarial consistency and the uniqueness of the adversarial bayes classifier
title_fullStr Adversarial consistency and the uniqueness of the adversarial bayes classifier
title_full_unstemmed Adversarial consistency and the uniqueness of the adversarial bayes classifier
title_short Adversarial consistency and the uniqueness of the adversarial bayes classifier
title_sort adversarial consistency and the uniqueness of the adversarial bayes classifier
topic Statistics theory
optimization
62A99
65K99
url https://www.cambridge.org/core/product/identifier/S0956792525000038/type/journal_article
work_keys_str_mv AT nataliesfrank adversarialconsistencyandtheuniquenessoftheadversarialbayesclassifier