Modeling Adversarial Noise for Adversarial Training

Dawei Zhou, Nannan Wang, Bo Han, Tongliang Liu
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:27353-27366, 2022.

Abstract

Deep neural networks have been demonstrated to be vulnerable to adversarial noise, promoting the development of defense against adversarial attacks. Motivated by the fact that adversarial noise contains well-generalizing features and that the relationship between adversarial data and natural data can help infer natural data and make reliable predictions, in this paper, we study to model adversarial noise by learning the transition relationship between adversarial labels (i.e. the flipped labels used to generate adversarial data) and natural labels (i.e. the ground truth labels of the natural data). Specifically, we introduce an instance-dependent transition matrix to relate adversarial labels and natural labels, which can be seamlessly embedded with the target model (enabling us to model stronger adaptive adversarial noise). Empirical evaluations demonstrate that our method could effectively improve adversarial accuracy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v162-zhou22k, title = {Modeling Adversarial Noise for Adversarial Training}, author = {Zhou, Dawei and Wang, Nannan and Han, Bo and Liu, Tongliang}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {27353--27366}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/zhou22k/zhou22k.pdf}, url = {https://proceedings.mlr.press/v162/zhou22k.html}, abstract = {Deep neural networks have been demonstrated to be vulnerable to adversarial noise, promoting the development of defense against adversarial attacks. Motivated by the fact that adversarial noise contains well-generalizing features and that the relationship between adversarial data and natural data can help infer natural data and make reliable predictions, in this paper, we study to model adversarial noise by learning the transition relationship between adversarial labels (i.e. the flipped labels used to generate adversarial data) and natural labels (i.e. the ground truth labels of the natural data). Specifically, we introduce an instance-dependent transition matrix to relate adversarial labels and natural labels, which can be seamlessly embedded with the target model (enabling us to model stronger adaptive adversarial noise). Empirical evaluations demonstrate that our method could effectively improve adversarial accuracy.} }
Endnote
%0 Conference Paper %T Modeling Adversarial Noise for Adversarial Training %A Dawei Zhou %A Nannan Wang %A Bo Han %A Tongliang Liu %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-zhou22k %I PMLR %P 27353--27366 %U https://proceedings.mlr.press/v162/zhou22k.html %V 162 %X Deep neural networks have been demonstrated to be vulnerable to adversarial noise, promoting the development of defense against adversarial attacks. Motivated by the fact that adversarial noise contains well-generalizing features and that the relationship between adversarial data and natural data can help infer natural data and make reliable predictions, in this paper, we study to model adversarial noise by learning the transition relationship between adversarial labels (i.e. the flipped labels used to generate adversarial data) and natural labels (i.e. the ground truth labels of the natural data). Specifically, we introduce an instance-dependent transition matrix to relate adversarial labels and natural labels, which can be seamlessly embedded with the target model (enabling us to model stronger adaptive adversarial noise). Empirical evaluations demonstrate that our method could effectively improve adversarial accuracy.
APA
Zhou, D., Wang, N., Han, B. & Liu, T.. (2022). Modeling Adversarial Noise for Adversarial Training. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:27353-27366 Available from https://proceedings.mlr.press/v162/zhou22k.html.

Related Material