Thwarting Adversarial Examples: An $L_0$-RobustSparse Fourier Transform

Authors: Mitali Bafna, Jack Murtagh, Nikhil Vyas
Published: 2018-12-12

Source: https://arxiv.org/abs/1812.05013

PDF: https://arxiv.org/pdf/1812.05013

AIにより推定されたラベル

ロバスト性 収束性分析 敵対的サンプルの検知

※ こちらのラベルはAIによって自動的に追加されました。そのため、正確でないことがあります。
詳細は文献データベースについてをご覧ください。

Abstract

We give a new algorithm for approximating the Discrete Fourier transform of an approximately sparse signal that has been corrupted by worst-case L₀ noise, namely a bounded number of coordinates of the signal have been corrupted arbitrarily. Our techniques generalize to a wide range of linear transformations that are used in data analysis such as the Discrete Cosine and Sine transforms, the Hadamard transform, and their high-dimensional analogs. We use our algorithm to successfully defend against well known L₀ adversaries in the setting of image classification. We give experimental results on the Jacobian-based Saliency Map Attack (JSMA) and the Carlini Wagner (CW) L₀ attack on the MNIST and Fashion-MNIST datasets as well as the Adversarial Patch on the ImageNet dataset.