Enregistré dans:
Détails bibliographiques
Auteurs principaux: Tamayo-Rousseau, Camilo, Zhao, Yunjia, Zhang, Yiqun, Balestriero, Randall
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2507.20453
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915482548830208
author Tamayo-Rousseau, Camilo
Zhao, Yunjia
Zhang, Yiqun
Balestriero, Randall
author_facet Tamayo-Rousseau, Camilo
Zhao, Yunjia
Zhang, Yiqun
Balestriero, Randall
contents Self-attention mechanisms are foundational to Transformer architectures, supporting their impressive success in a wide range of tasks. While there are many self-attention variants, their robustness to noise and spurious correlations has not been well studied. This study evaluates Softmax, Sigmoid, Linear, Doubly Stochastic, and Cosine attention within Vision Transformers under different data corruption scenarios. Through testing across the CIFAR-10, CIFAR-100, and Imagenette datasets, we show that Doubly Stochastic attention is the most robust. It consistently outperformed the next best mechanism by $0.1\%-5.1\%$ when training data, or both training and testing data, were corrupted. Our findings inform self-attention selection in contexts with imperfect data. The code used is available at https://github.com/ctamayor/NeurIPS-Robustness-ViT.
format Preprint
id arxiv_https___arxiv_org_abs_2507_20453
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Your Attention Matters: to Improve Model Robustness to Noise and Spurious Correlations
Tamayo-Rousseau, Camilo
Zhao, Yunjia
Zhang, Yiqun
Balestriero, Randall
Machine Learning
Self-attention mechanisms are foundational to Transformer architectures, supporting their impressive success in a wide range of tasks. While there are many self-attention variants, their robustness to noise and spurious correlations has not been well studied. This study evaluates Softmax, Sigmoid, Linear, Doubly Stochastic, and Cosine attention within Vision Transformers under different data corruption scenarios. Through testing across the CIFAR-10, CIFAR-100, and Imagenette datasets, we show that Doubly Stochastic attention is the most robust. It consistently outperformed the next best mechanism by $0.1\%-5.1\%$ when training data, or both training and testing data, were corrupted. Our findings inform self-attention selection in contexts with imperfect data. The code used is available at https://github.com/ctamayor/NeurIPS-Robustness-ViT.
title Your Attention Matters: to Improve Model Robustness to Noise and Spurious Correlations
topic Machine Learning
url https://arxiv.org/abs/2507.20453