Bounding data reconstruction attacks with the hypothesis testing interpretation of differential privacy
This work provides a direct, computable bound for data reconstruction attacks, addressing a specific need in privacy-preserving machine learning, but it is incremental as it builds on existing concepts.
The paper tackled the problem of bounding data reconstruction attacks by connecting hypothesis testing differential privacy to Reconstruction Robustness, deriving closed-form bounds for Laplace and Gaussian mechanisms and their subsampled variants.
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.