论文标题
通过混合和有针对性的标签对抗训练来解决神经网络的鲁棒性
Addressing Neural Network Robustness with Mixup and Targeted Labeling Adversarial Training
论文作者
论文摘要
尽管表现出色，但人工神经网络对于大多数工业应用还不够可靠。它们对噪音，旋转，模糊和对抗性例子很敏感。有必要建立防御能力，以防止各种扰动，涵盖最传统的常见腐败和对抗性例子。我们提出了一种称为M-TLAT的新数据增强策略，旨在在广义上解决鲁棒性。我们的方法结合了混合增强和一种称为靶向标签对抗训练（TLAT）的新的对抗训练算法。 TLAT的想法是用地面真实标签插入对抗示例的目标标签。我们表明，M-TLAT可以提高图像分类器对19个常见腐败和五次对抗性攻击的鲁棒性，而无需降低清洁样品的准确性。
Despite their performance, Artificial Neural Networks are not reliable enough for most of industrial applications. They are sensitive to noises, rotations, blurs and adversarial examples. There is a need to build defenses that protect against a wide range of perturbations, covering the most traditional common corruptions and adversarial examples. We propose a new data augmentation strategy called M-TLAT and designed to address robustness in a broad sense. Our approach combines the Mixup augmentation and a new adversarial training algorithm called Targeted Labeling Adversarial Training (TLAT). The idea of TLAT is to interpolate the target labels of adversarial examples with the ground-truth labels. We show that M-TLAT can increase the robustness of image classifiers towards nineteen common corruptions and five adversarial attacks, without reducing the accuracy on clean samples.
