AISTATS 2021poster78 citations

Federated f-Differential Privacy

Qinqing Zheng, Shuxiao Chen, Qi Long, Weijie Su

Abstract

Federated learning (FL) is a training paradigm where the clients collaboratively learn models by repeatedly sharing information without compromising much on the privacy of their local sensitive data. In this paper, we introduce \emph{federated $f$-differential privacy}, a new notion specifically tailored to the federated setting, based on the framework of Gaussian differential privacy. Federated $f$-differential privacy operates on \emph{record level}: it provides the privacy guarantee on each individual record of one client’s data against adversaries. We then propose a generic private federated learning framework \fedsync that accommodates a large family of state-of-the-art FL algorithms, which provably achieves {federated $f$-differential privacy}. Finally, we empirically demonstrate the trade-off between privacy guarantee and prediction performance for models trained by \fedsync in computer vision tasks.

BibTeX
@InProceedings{pmlr-v130-zheng21a,
  title = 	 { Federated f-Differential Privacy },
  author =       {Zheng, Qinqing and Chen, Shuxiao and Long, Qi and Su, Weijie},
  booktitle = 	 {Proceedings of The 24th International Conference on Artificial Intelligence and Statistics},
  pages = 	 {2251--2259},
  year = 	 {2021},
  editor = 	 {Banerjee, Arindam and Fukumizu, Kenji},
  volume = 	 {130},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {13--15 Apr},
  publisher =    {PMLR},
  pdf = 	 {http://proceedings.mlr.press/v130/zheng21a/zheng21a.pdf},
  url = 	 {https://proceedings.mlr.press/v130/zheng21a.html},
  abstract = 	 { Federated learning (FL) is a training paradigm where the clients collaboratively learn models by repeatedly sharing information without compromising much on the privacy of their local sensitive data. In this paper, we introduce \emph{federated $f$-differential privacy}, a new notion specifically tailored to the federated setting, based on the framework of Gaussian differential privacy. Federated $f$-differential privacy operates on \emph{record level}: it provides the privacy guarantee on each individual record of one client’s data against adversaries. We then propose a generic private federated learning framework \fedsync that accommodates a large family of state-of-the-art FL algorithms, which provably achieves {federated $f$-differential privacy}. Finally, we empirically demonstrate the trade-off between privacy guarantee and prediction performance for models trained by \fedsync in computer vision tasks. }
}