Indiscriminate data poisoning attacks are quite effective against supervised
learning. However, not much is known about their impact on unsupervised
contrastive learning (CL). This paper is the first to consider indiscriminate
poisoning attacks of contrastive learning. We propose Contrastive Poisoning
(CP), the first effective such attack on CL. We empirically show that
Contrastive Poisoning, not only drastically reduces the performance of CL
algorithms, but also attacks supervised learning models, making it the most
generalizable indiscriminate poisoning attack. We also show that CL algorithms
with a momentum encoder are more robust to indiscriminate poisoning, and
propose a new countermeasure based on matrix completion. Code is available at:
this https URL.

本文针对无监督对比学习中的不分靶标数据污染攻击进行了研究。该研究首次提出了 Contrastive Poisoning，这是一种破坏性最强、最具一般性的不分靶标攻击方式。同时针对这种攻击，提出了一种基于矩阵填充的新型对策。实验证明，强调数据惯性的编码器可以更加有效地抵御这种攻击。