SAIF: Sparse Adversarial and Imperceptible Attack Framework
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2022
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909787257569280 |
|---|---|
| author | Imtiaz, Tooba Kohler, Morgan Miller, Jared Wang, Zifeng Eskandar, Masih Sznaier, Mario Camps, Octavia Dy, Jennifer |
| author_facet | Imtiaz, Tooba Kohler, Morgan Miller, Jared Wang, Zifeng Eskandar, Masih Sznaier, Mario Camps, Octavia Dy, Jennifer |
| contents | Adversarial attacks hamper the decision-making ability of neural networks by perturbing the input signal. The addition of calculated small distortion to images, for instance, can deceive a well-trained image classification network. In this work, we propose a novel attack technique called Sparse Adversarial and Interpretable Attack Framework (SAIF). Specifically, we design imperceptible attacks that contain low-magnitude perturbations at a small number of pixels and leverage these sparse attacks to reveal the vulnerability of classifiers. We use the Frank-Wolfe (conditional gradient) algorithm to simultaneously optimize the attack perturbations for bounded magnitude and sparsity with $O(1/\sqrt{T})$ convergence. Empirical results show that SAIF computes highly imperceptible and interpretable adversarial examples, and outperforms state-of-the-art sparse attack methods on the ImageNet dataset. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2212_07495 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | SAIF: Sparse Adversarial and Imperceptible Attack Framework Imtiaz, Tooba Kohler, Morgan Miller, Jared Wang, Zifeng Eskandar, Masih Sznaier, Mario Camps, Octavia Dy, Jennifer Computer Vision and Pattern Recognition Adversarial attacks hamper the decision-making ability of neural networks by perturbing the input signal. The addition of calculated small distortion to images, for instance, can deceive a well-trained image classification network. In this work, we propose a novel attack technique called Sparse Adversarial and Interpretable Attack Framework (SAIF). Specifically, we design imperceptible attacks that contain low-magnitude perturbations at a small number of pixels and leverage these sparse attacks to reveal the vulnerability of classifiers. We use the Frank-Wolfe (conditional gradient) algorithm to simultaneously optimize the attack perturbations for bounded magnitude and sparsity with $O(1/\sqrt{T})$ convergence. Empirical results show that SAIF computes highly imperceptible and interpretable adversarial examples, and outperforms state-of-the-art sparse attack methods on the ImageNet dataset. |
| title | SAIF: Sparse Adversarial and Imperceptible Attack Framework |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2212.07495 |