Provably Learning Diverse Features in Multi-View Data with Midpoint Mixup

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chidambaram, Muthu, Wang, Xiang, Wu, Chenwei, Ge, Rong
Natura: Preprint
Pubblicazione: 2022
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916467649282048
author Chidambaram, Muthu
Wang, Xiang
Wu, Chenwei
Ge, Rong
author_facet Chidambaram, Muthu
Wang, Xiang
Wu, Chenwei
Ge, Rong
contents Mixup is a data augmentation technique that relies on training using random convex combinations of data points and their labels. In recent years, Mixup has become a standard primitive used in the training of state-of-the-art image classification models due to its demonstrated benefits over empirical risk minimization with regards to generalization and robustness. In this work, we try to explain some of this success from a feature learning perspective. We focus our attention on classification problems in which each class may have multiple associated features (or views) that can be used to predict the class correctly. Our main theoretical results demonstrate that, for a non-trivial class of data distributions with two features per class, training a 2-layer convolutional network using empirical risk minimization can lead to learning only one feature for almost all classes while training with a specific instantiation of Mixup succeeds in learning both features for every class. We also show empirically that these theoretical insights extend to the practical settings of image benchmarks modified to have multiple features.
format Preprint
id arxiv_https___arxiv_org_abs_2210_13512
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Provably Learning Diverse Features in Multi-View Data with Midpoint Mixup
Chidambaram, Muthu
Wang, Xiang
Wu, Chenwei
Ge, Rong
Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
Optimization and Control
Mixup is a data augmentation technique that relies on training using random convex combinations of data points and their labels. In recent years, Mixup has become a standard primitive used in the training of state-of-the-art image classification models due to its demonstrated benefits over empirical risk minimization with regards to generalization and robustness. In this work, we try to explain some of this success from a feature learning perspective. We focus our attention on classification problems in which each class may have multiple associated features (or views) that can be used to predict the class correctly. Our main theoretical results demonstrate that, for a non-trivial class of data distributions with two features per class, training a 2-layer convolutional network using empirical risk minimization can lead to learning only one feature for almost all classes while training with a specific instantiation of Mixup succeeds in learning both features for every class. We also show empirically that these theoretical insights extend to the practical settings of image benchmarks modified to have multiple features.
title Provably Learning Diverse Features in Multi-View Data with Midpoint Mixup
topic Machine Learning
Artificial Intelligence
Computer Vision and Pattern Recognition
Optimization and Control
url https://arxiv.org/abs/2210.13512