We demonstrate, theoretically and empirically, that adversarial robustness can significantly benefit from semisupervised learning. Theoretically, we revisit the simple Gaussian model of Schmidt et al. that shows a sample complexity gap between standard and robust classification. We prove that this gap does not pertain to labels: a simple semisupervised learning procedure (self-training) achieves robust accuracy using the same number of labels required for standard accuracy. Empirically, we augment CIFAR-10 with 500K unlabeled images sourced from 80 Million Tiny Images and use robust self-training to outperform state-of-the-art robust accuracies by over 5 points in (i) $\ell_\infty$ robustness against several strong attacks via adversarial training and (ii) certified $\ell_2$ and $\ell_\infty$ robustness via randomized smoothing. On SVHN, adding the dataset's own extra training set with the labels removed provides gains of 4 to 10 points, within 1 point of the gain from using the extra labels as well.

通过理论和实验，我们证明了半监督学习可以显著提高对抗性鲁棒性，实验结果表明在 CIFAR-10 上使用 500k 未标记图像，使用自我训练方法可以超过最先进的对抗性鲁棒的准确度。在 SVHN 上，使用模型自身的额外的训练集可以提高 4 至 10 个百分点，与使用额外标签的提高量相差不大。

无标签数据提高对抗鲁棒性