Subsampled Renyi Differential Privacy and Analytical Moments Accountant

Abstract

We study the problem of subsampling in differential privacy (DP), a question that is the centerpiece behind many successful differentially private machine learning algorithms. Specifically, we provide a tight upper bound on the Renyi Differential Privacy (RDP) [Mironov 2017] parameters for algorithms that: (1) subsample the dataset, and then (2) applies a randomized mechanism M to the subsample, in terms of the RDP parameters of M and the subsampling probability parameter. Our results generalize the moments accounting technique, developed by [Abadi et al. 2016] for the Gaussian mechanism, to any subsampled RDP mechanism.

Cite

Text

Wang et al. "Subsampled Renyi Differential Privacy and Analytical Moments Accountant." Artificial Intelligence and Statistics, 2019.

Markdown

[Wang et al. "Subsampled Renyi Differential Privacy and Analytical Moments Accountant." Artificial Intelligence and Statistics, 2019.](https://mlanthology.org/aistats/2019/wang2019aistats-subsampled/)

BibTeX

@inproceedings{wang2019aistats-subsampled,
  title     = {{Subsampled Renyi Differential Privacy and Analytical Moments Accountant}},
  author    = {Wang, Yu-Xiang and Balle, Borja and Kasiviswanathan, Shiva Prasad},
  booktitle = {Artificial Intelligence and Statistics},
  year      = {2019},
  pages     = {1226-1235},
  volume    = {89},
  url       = {https://mlanthology.org/aistats/2019/wang2019aistats-subsampled/}
}