Membership Inference Attacks From First Principles

Nicholas Carlini, Steve Chien, Milad Nasr, Shuang Song, Andreas Terzis, and Florian Tramèr   (alphabetical author ordering)

IEEE Symposium on Security and Privacy (S&P) 2022



Abstract

A membership inference attack allows an adversary to query a trained machine learning model to predict whether or not a particular example was contained in the model’s training dataset. These attacks are currently evaluated using average-case “accuracy” metrics that fail to characterize whether the attack can confidently identify any members of the training set. We argue that attacks should instead be evaluated by computing their true-positive rate at low (e.g., <0.1%) false-positive rates, and find most prior attacks perform poorly when evaluated in this way. To address this we develop a Likelihood Ratio Attack (LiRA) that carefully combines multiple ideas from the literature. Our attack is 10x more powerful at low false-positive rates, and also strictly dominates prior attacks on existing metrics.


BibTeX
@inproceedings{CCNS+22,
  author   =   {Carlini, Nicholas and Chien, Steve and Nasr, Milad and Song, Shuang and Terzis, Andreas and Tram{\`e}r, Florian},
  title   =   {Membership Inference Attacks From First Principles},
  booktitle   =   {IEEE Symposium on Security and Privacy (S\&P)},
  year   =   {2022},
  howpublished   =   {arXiv preprint arXiv:2112.03570},
  url   =   {https://arxiv.org/abs/2112.03570}
}