Subsampled Rényi Differential Privacy and Analytical Moments Accountant

Authors: Yu-Xiang Wang, Borja Balle, Shiva Kasiviswanathan
Published: 2018-07-31 | Updated: 2018-12-04

Source: https://arxiv.org/abs/1808.00087

PDF: https://arxiv.org/pdf/1808.00087

Labels Predicted by AI

プライバシー評価 差分プライバシー RDPの特性

Please note that these labels were automatically added by AI. Therefore, they may not be entirely accurate.
For more details, please see the About the Literature Database page.

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 Rényi 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.