A Margin-based Multiclass Generalization Bound via Geometric Complexity

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Munn, Michael, Dherin, Benoit, Gonzalvo, Javier
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913368147755008
author Munn, Michael
Dherin, Benoit
Gonzalvo, Javier
author_facet Munn, Michael
Dherin, Benoit
Gonzalvo, Javier
contents There has been considerable effort to better understand the generalization capabilities of deep neural networks both as a means to unlock a theoretical understanding of their success as well as providing directions for further improvements. In this paper, we investigate margin-based multiclass generalization bounds for neural networks which rely on a recent complexity measure, the geometric complexity, developed for neural networks. We derive a new upper bound on the generalization error which scales with the margin-normalized geometric complexity of the network and which holds for a broad family of data distributions and model classes. Our generalization bound is empirically investigated for a ResNet-18 model trained with SGD on the CIFAR-10 and CIFAR-100 datasets with both original and random labels.
format Preprint
id arxiv_https___arxiv_org_abs_2405_18590
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Margin-based Multiclass Generalization Bound via Geometric Complexity
Munn, Michael
Dherin, Benoit
Gonzalvo, Javier
Machine Learning
There has been considerable effort to better understand the generalization capabilities of deep neural networks both as a means to unlock a theoretical understanding of their success as well as providing directions for further improvements. In this paper, we investigate margin-based multiclass generalization bounds for neural networks which rely on a recent complexity measure, the geometric complexity, developed for neural networks. We derive a new upper bound on the generalization error which scales with the margin-normalized geometric complexity of the network and which holds for a broad family of data distributions and model classes. Our generalization bound is empirically investigated for a ResNet-18 model trained with SGD on the CIFAR-10 and CIFAR-100 datasets with both original and random labels.
title A Margin-based Multiclass Generalization Bound via Geometric Complexity
topic Machine Learning
url https://arxiv.org/abs/2405.18590