Work

Bounding data reconstruction attacks with the hypothesis testing interpretation of differential privacy

Year: 2023

Type: preprint

Abstract: We explore Reconstruction Robustness (ReRo), which was recently proposed as an upper bound on the success of data reconstruction attacks against machine learning models. Previous research has demonstrated that differential privacy (DP) mechanisms also provide ReRo, but so far, only asymptotic Monte Carlo estimates of a tight ReRo bound have been shown. Directly computable ReRo bounds for general DP mechanisms are thus desirable. In this work, we establish a connection between hypothesis testing DP and ReRo and derive closed-form, analytic or numerical ReRo bounds for the Laplace and Gaussian mechanisms and their subsampled variants.

Source: arXiv (Cornell University)

DOI: 10.48550/arxiv.2307.03928

Authors: Kaissis, Georgios, Hayes, Jamie, Ziller, Alexander, Rueckert, Daniel

Institutions: Google DeepMind (United Kingdom), Technical University of Munich

Language: English

Topic: Privacy-Preserving Technologies in Data

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W4383987414

Cites:

Cited by: 1

Related to: 10