Contrastive Learning first extracts features from unlabeled data, followed by linear probing with labeled data. Adversarial Contrastive Learning (ACL) integrates Adversarial Training into the first phase to enhance feature robustness against attacks in the probing phase. While ACL has shown strong empirical results, its theoretical understanding remains limited. Furthermore, while a fair amount of theoretical works analyze how the unsupervised loss can support the supervised loss in the probing phase, none has examined its role to the robust supervised loss. To fill this gap, our work develops rigorous theories to identify which components in the unsupervised training can help improve the robust supervised loss. Specifically, besides the adversarial contrastive loss, we reveal that the benign one, along with a global divergence between benign and adversarial examples can also improve robustness. Proper experiments are conducted to justify our findings.

翻译：对比学习首先从无标签数据中提取特征，随后利用有标签数据进行线性探测。对抗对比学习将对抗训练整合到第一阶段，以增强特征在探测阶段面对攻击的鲁棒性。尽管对抗对比学习已展现出强大的实证结果，但其理论理解仍较为有限。此外，尽管已有相当数量的理论研究分析了无监督损失如何支持探测阶段的有监督损失，但尚未有研究探讨其对鲁棒有监督损失的作用。为填补这一空白，本文建立了严格的理论框架，以识别无监督训练中哪些组件有助于提升鲁棒有监督损失。具体而言，除了对抗对比损失外，我们发现良性对比损失以及良性样本与对抗样本之间的全局散度同样能增强鲁棒性。我们通过适当的实验验证了这些发现。