Improving the Robustness of Deep Neural Networks via Adversarial Training with Triplet Loss

Authors: Pengcheng Li, Jinfeng Yi, Bowen Zhou, Lijun Zhang
Published: 2019-05-28

Source: https://arxiv.org/abs/1905.11713

PDF: https://arxiv.org/pdf/1905.11713

AIにより推定されたラベル

敵対的学習 敵対的サンプル ポイズニング

※ 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

Recent studies have highlighted that deep neural networks (DNNs) are vulnerable to adversarial examples. In this paper, we improve the robustness of DNNs by utilizing techniques of Distance Metric Learning. Specifically, we incorporate Triplet Loss, one of the most popular Distance Metric Learning methods, into the framework of adversarial training. Our proposed algorithm, Adversarial Training with Triplet Loss (AT²L), substitutes the adversarial example against the current model for the anchor of triplet loss to effectively smooth the classification boundary. Furthermore, we propose an ensemble version of AT²L, which aggregates different attack methods and model structures for better defense effects. Our empirical studies verify that the proposed approach can significantly improve the robustness of DNNs without sacrificing accuracy. Finally, we demonstrate that our specially designed triplet loss can also be used as a regularization term to enhance other defense methods.