Masking and Mixing Adversarial Training

Hiroki Adachi, Tsubasa Hirakawa, Takayoshi Yamashita, Hironobu Fujiyoshi, Yasunori Ishii, Kazuki Kozuka

2023

Abstract

While convolutional neural networks (CNNs) have achieved excellent performances in various computer vision tasks, they often misclassify with malicious samples, a.k.a. adversarial examples. Adversarial training is a popular and straightforward technique to defend against the threat of adversarial examples. Unfortunately, CNNs must sacrifice the accuracy of standard samples to improve robustness against adversarial examples when adversarial training is used. In this work, we propose Masking and Mixing Adversarial Training (M2 AT) to mitigate the trade-off between accuracy and robustness. We focus on creating diverse adversarial examples during training. Specifically, our approach consists of two processes: 1) masking a perturbation with a binary mask and 2) mixing two partially perturbed images. Experimental results on CIFAR-10 dataset demonstrate that our method achieves better robustness against several adversarial attacks than previous methods.

Paper Citation

in Harvard Style

Adachi H., Hirakawa T., Yamashita T., Fujiyoshi H., Ishii Y. and Kozuka K. (2023). Masking and Mixing Adversarial Training. In Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 4: VISAPP; ISBN 978-989-758-634-7, SciTePress, pages 74-82. DOI: 10.5220/0011653300003417

in Bibtex Style

@conference{visapp23,
author={Hiroki Adachi and Tsubasa Hirakawa and Takayoshi Yamashita and Hironobu Fujiyoshi and Yasunori Ishii and Kazuki Kozuka},
title={Masking and Mixing Adversarial Training},
booktitle={Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 4: VISAPP},
year={2023},
pages={74-82},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011653300003417},
isbn={978-989-758-634-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2023) - Volume 4: VISAPP
TI - Masking and Mixing Adversarial Training
SN - 978-989-758-634-7
AU - Adachi H.
AU - Hirakawa T.
AU - Yamashita T.
AU - Fujiyoshi H.
AU - Ishii Y.
AU - Kozuka K.
PY - 2023
SP - 74
EP - 82
DO - 10.5220/0011653300003417
PB - SciTePress