Semantic Scholar Open Access 2023 5 sitasi

Revisiting Image Classifier Training for Improved Certified Robust Defense against Adversarial Patches

Aniruddha Saha Shuhua Yu Arash Norouzzadeh Wan-Yi Lin Chaithanya Kumar Mummadi

Abstrak

Certifiably robust defenses against adversarial patches for image classifiers ensure correct prediction against any changes to a constrained neighborhood of pixels. PatchCleanser arXiv:2108.09135 [cs.CV], the state-of-the-art certified defense, uses a double-masking strategy for robust classification. The success of this strategy relies heavily on the model's invariance to image pixel masking. In this paper, we take a closer look at model training schemes to improve this invariance. Instead of using Random Cutout arXiv:1708.04552v2 [cs.CV] augmentations like PatchCleanser, we introduce the notion of worst-case masking, i.e., selecting masked images which maximize classification loss. However, finding worst-case masks requires an exhaustive search, which might be prohibitively expensive to do on-the-fly during training. To solve this problem, we propose a two-round greedy masking strategy (Greedy Cutout) which finds an approximate worst-case mask location with much less compute. We show that the models trained with our Greedy Cutout improves certified robust accuracy over Random Cutout in PatchCleanser across a range of datasets and architectures. Certified robust accuracy on ImageNet with a ViT-B16-224 model increases from 58.1\% to 62.3\% against a 3\% square patch applied anywhere on the image.

Topik & Kata Kunci

Penulis (5)

A

Aniruddha Saha

S

Shuhua Yu

A

Arash Norouzzadeh

W

Wan-Yi Lin

C

Chaithanya Kumar Mummadi

Format Sitasi

Saha, A., Yu, S., Norouzzadeh, A., Lin, W., Mummadi, C.K. (2023). Revisiting Image Classifier Training for Improved Certified Robust Defense against Adversarial Patches. https://doi.org/10.48550/arXiv.2306.12610

Akses Cepat

Lihat di Sumber doi.org/10.48550/arXiv.2306.12610
Informasi Jurnal
Tahun Terbit
2023
Bahasa
en
Total Sitasi
Sumber Database
Semantic Scholar
DOI
10.48550/arXiv.2306.12610
Akses
Open Access ✓