Fast Training of Provably Robust Neural Networks by SingleProp

Authors: Akhilan Boopathy, Tsui-Wei Weng, Sijia Liu, Pin-Yu Chen, Gaoyuan Zhang, Luca Daniel
Published: 2021-02-01

Source: https://arxiv.org/abs/2102.01208

PDF: https://arxiv.org/pdf/2102.01208

Labels Predicted by AI

バックドア攻撃 敵対的訓練 モデル性能評価

Please note that these labels were automatically added by AI. Therefore, they may not be entirely accurate.
For more details, please see the About the Literature Database page.

Abstract

Recent works have developed several methods of defending neural networks against adversarial attacks with certified guarantees. However, these techniques can be computationally costly due to the use of certification during training. We develop a new regularizer that is both more efficient than existing certified defenses, requiring only one additional forward propagation through a network, and can be used to train networks with similar certified accuracy. Through experiments on MNIST and CIFAR-10 we demonstrate improvements in training speed and comparable certified accuracy compared to state-of-the-art certified defenses.