Understanding and Improving Fast Adversarial Training

By Maksym Andriushchenko, Nicolas Flammarion. Neural Information Processing Systems, 2020

353 citations recorded by the source corpus.

A recent line of work focused on making adversarial training computationally efficient for deep learning models. In particular, Wong et al. (2020) showed that $\ell_\infty$-adversarial training with fast gradient sign method (FGSM) can fail due to a phenomenon called "catastrophic overfitting", when the model quickly loses its robustness over a single epoch of training. We show that adding a random step to FGSM, as proposed in Wong et al. (2020), does not prevent catastrophic overfitting, and that randomness is not important per se -- its main role being simply to reduce the magnitude of the perturbation. Moreover, we show that catastrophic overfitting is not inherent to deep and overparametrized networks, but can occur in a single-layer convolutional network with a few filters. In an extreme case, even a single filter can make the network highly non-linear locally, which is the main reason why FGSM training fails. Based on this observation, we propose a new regularization method, GradAlign, that prevents catastrophic overfitting by explicitly maximizing the gradient alignment inside the perturbation set and improves the quality of the FGSM solution. As a result, GradAlign allows to successfully apply FGSM training also for larger $\ell_\infty$-perturbations and reduce the gap to multi-step adversarial training. The code of our experiments is available at this https URL.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/24cf86a418c9471e8001961c87697c825f0bba8f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Cross-Entropy Loss Functions: Theoretical Analysis and Applications (2023, background): “…et al., 2020a; Ding et al., 2022; Cheng et al., 2022), efficient adversarial training (Shafahi et al., 2019; Zhang et al., 2019a; Wong et al., 2020; Andriushchenko & Flammarion, 2020), ensemble techniques (Tramèr et al., 2018; Pang et al., 2019; Yang et al., 2020; Guo et al., 2022), unlabeled data…”
Relating Adversarially Robust Generalization to Flat Minima (2021, neutral): “Originally proposed in different variants in [56, 40, 24], it received considerable attention in [37, 16] and has been extended in various ways: [33, 7, 2] utilize interpolated or unlabeled examples, [57, 38] achieve robustness against multiple threat models, [54, 32, 64] augment AT with a reject op”
Recent Advances in Adversarial Training for Adversarial Robustness (2021, background): “However, it is found [Andriushchenko and Flammarion, 2020] that the above fast training methods [Shafahi et al. , 2019; Wong et al. , 2020] suffer from catastrophic overﬁtting like FGSM-AT.”
Better Diffusion Models Further Improve Adversarial Training (2023, background): “Some following studies address the side effects (e.g., catastrophic overfitting) induced by these fast AT approaches (Andriushchenko & Flammarion, 2020; Li et al., 2020).”
Uncovering the Limits of Adversarial Training against Norm-Bounded Adversarial Examples (2020, background): “Since Madry et al. (2017), various modiﬁcations to their algorithm have been proposed (Xie et al., 2018; Pang et al., 2020b; Huang et al., 2020; Qin et al., 2019; Zoran et al., 2019; Andriushchenko & Flammarion, 2020).”
Advances in adversarial attacks and defenses in computer vision: A survey (2021, background): “Their improvement of adding randomization with FGSM-based adversarial training is, however, contradicted to an extend by [358].”
Bag of Tricks for Adversarial Training (2020, methodology): “The following works try to solve the side effects (e.g., catastrophic overﬁtting) caused by these fast AT methods (Andriushchenko & Flammarion, 2020; Li et al., 2020).”