Differentially Private Synthetic Data Generation for Relational Databases

Abstract

Existing differentially private (DP) synthetic data generation mechanisms typically assume a single-source table. In practice, data is often distributed across multiple tables with relationships across tables. This study presents the first-of-its-kind algorithm that can be combined with \emph{any} existing DP mechanisms to generate synthetic relational databases. The algorithm iteratively refines the relationship between individual synthetic tables to minimize their approximation errors in terms of low-order marginal distributions while maintaining referential integrity; consequently eliminates the need to flatten a relational database into a master table (saving space), operates efficiently (saving time), and scales effectively to high-dimensional data. We provide both DP and theoretical utility guarantees for our algorithm. Through numerical experiments on real-world datasets, we demonstrate the effectiveness of our method in preserving fidelity to the original data.