GEM+: Scalable State-of-the-Art Private Synthetic Data with Generator Networks

TL;DR

GEM+ introduces a scalable, differentially private synthetic data generation method that combines adaptive measurement with generator neural networks, outperforming previous approaches on high-dimensional datasets.

Contribution

This work presents GEM+, a novel method integrating AIM's adaptive framework with GEM's scalable neural network generators for high-dimensional private synthetic data.

Findings

GEM+ outperforms AIM in data utility and scalability.

GEM+ effectively handles datasets with over a hundred columns.

GEM+ reduces computational overhead compared to previous methods.

Abstract

State-of-the-art differentially private synthetic tabular data has been defined by adaptive 'select-measure-generate' frameworks, exemplified by methods like AIM. These approaches iteratively measure low-order noisy marginals and fit graphical models to produce synthetic data, enabling systematic optimisation of data quality under privacy constraints. Graphical models, however, are inefficient for high-dimensional data because they require substantial memory and must be retrained from scratch whenever the graph structure changes, leading to significant computational overhead. Recent methods, like GEM, overcome these limitations by using generator neural networks for improved scalability. However, empirical comparisons have mostly focused on small datasets, limiting real-world applicability. In this work, we introduce GEM+, which integrates AIM's adaptive measurement framework with GEM's scalable generator network. Our experiments show that GEM+ outperforms AIM in both utility and scalability, delivering state-of-the-art results while efficiently handling datasets with over a hundred columns, where AIM fails due to memory and computational overheads.