Vaader Seminar: Detect and Defense Against Adversarial Examples in Deep Learning using Natural Scene Statistics and Adaptive Denoising - Wassim Hamidouche

Despite the enormous performance of deep neural networks (DNNs), recent studies have shown their vulnerability to adversarial examples (AEs), i.e., care-fully perturbed inputs designed to fool the targeted DNN. Currently, the literature is rich with many effective attacks to craft such AEs. Meanwhile, many defenses strategies have been developed to mitigate this vulnerability. However, these latter showed their effectiveness against specific attacks and does not generalize well to different attacks. In this paper, we propose a framework for defending DNN classifier against adversarial samples. The proposed method is based on a two-stage framework involving a separate detector and a denoising block. The detector aims to detect AEs by characterizing them through the use of natural scene statistic (NSS), where we demonstrate that these statistical features are altered by the presence of adversarial perturbations. The denoiser is based on block matching3D (BM3D) filter fed by an optimum threshold value estimated by a convolutional neural network (CNN) toproject back the samples detected as AEs into their data manifold. We conducted a complete evaluation on three standard datasets namely MNIST, CIFAR-10 and Tiny-ImageNet. The experimental results show that the proposed defense method outperforms the state-of-the-art defense techniques by improving the robustness against a set of attacks under black-box, gray-box and white-box settings.

Mots clés : artificial intelligence deep fakes deep learning ietr machine learning vaader