Boosting Adversarial Attacks with Momentum

• DOI: 10.48550/arxiv.1710.06081
• Publication Date: 2017-01-01
• Authors: Yinpeng Dong; Fangzhou Liao; Tianyu Pang; Hang Su; Jun Zhu; Xiaolin Hu; Jianguo Li
• Keywords: Robustness; Boosting; Deep Neural Networks; Benchmark (surveying); Black box

Deep neural networks are vulnerable to adversarial examples, which poses security concerns on these algorithms due the potentially severe consequences. Adversarial attacks serve as an important surrogate evaluate robustness of deep learning models before they deployed. However, most existing can only fool a black-box model with low success rate. To address this issue, we propose broad class momentum-based iterative boost attacks. By integrating momentum term into process for attacks, our methods stabilize update directions and escape from poor local maxima during iterations, resulting in more transferable examples. further improve rates apply ensemble models, show that adversarially trained strong defense ability also We hope proposed will benchmark evaluating various methods. With method, won first places NIPS 2017 Non-targeted Attack Targeted competitions.