Abstract
Current deep neural networks (DNNs) are vulnerable to adversarial attacks, where adversarial perturbations to the inputs can change or manipulate classification. To defend against such attacks, an effective and popular approach, known as adversarial training (AT), has been shown to mitigate the negative impact of adversarial attacks by virtue of a min-max robust training method. While effective, it remains unclear whether it can successfully be adapted to the distributed learning context. The power of distributed optimization over multiple machines enables us to scale up robust training over large models and datasets. Spurred by that, we propose distributed adversarial training (DAT), a large-batch adversarial training framework implemented over multiple machines. We show that DAT is general, which supports training over labeled and unlabeled data, multiple types of attack generation methods, and gradient compression operations favored for distributed optimization. Theoretically, we provide, under standard conditions in the optimization theory, the convergence rate of DAT to the first-order stationary points in general non-convex settings. Empirically, we demonstrate that DAT either matches or outperforms state-of-the-art robust accuracies and achieves a graceful training speedup (e.g., on ResNet-50 under ImageNet). Codes are available at https://github.com/dat-2022/dat.
Original language | English (US) |
---|---|
Title of host publication | Proceedings of the 38th Conference on Uncertainty in Artificial Intelligence, UAI 2022 |
Publisher | Association For Uncertainty in Artificial Intelligence (AUAI) |
Pages | 2353-2363 |
Number of pages | 11 |
ISBN (Electronic) | 9781713863298 |
State | Published - 2022 |
Event | 38th Conference on Uncertainty in Artificial Intelligence, UAI 2022 - Eindhoven, Netherlands Duration: Aug 1 2022 → Aug 5 2022 |
Publication series
Name | Proceedings of the 38th Conference on Uncertainty in Artificial Intelligence, UAI 2022 |
---|
Conference
Conference | 38th Conference on Uncertainty in Artificial Intelligence, UAI 2022 |
---|---|
Country/Territory | Netherlands |
City | Eindhoven |
Period | 8/1/22 → 8/5/22 |
Bibliographical note
Funding Information:Y. Zhang and S. Liu are supported by the Cisco Research grant CG# 70614511. P. Khanduri and M. Hong are supported in part by the NSF grants 1910385 and 1727757. We also thank Dr. Cho-Jui Hsieh for the helpful discussion on early ideas of this paper.
Publisher Copyright:
© 2022 Proceedings of the 38th Conference on Uncertainty in Artificial Intelligence, UAI 2022. All right reserved.