Synthetic Data Outliers: Navigating Identity Disclosure

TL;DR

This paper investigates the privacy risks of synthetic data, especially outliers, revealing that re-identification is feasible and can be mitigated with differential privacy at the cost of data utility.

Contribution

It highlights the vulnerability of synthetic data outliers to re-identification attacks and evaluates privacy-preserving safeguards like differential privacy.

Findings

Outliers in synthetic data can be re-identified through linkage attacks.

Differential privacy can prevent re-identification.

Applying safeguards reduces data utility.

Abstract

Multiple synthetic data generation models have emerged, among which deep learning models have become the vanguard due to their ability to capture the underlying characteristics of the original data. However, the resemblance of the synthetic to the original data raises important questions on the protection of individuals' privacy. As synthetic data is perceived as a means to fully protect personal information, most current related work disregards the impact of re-identification risk. In particular, limited attention has been given to exploring outliers, despite their privacy relevance. In this work, we analyze the privacy of synthetic data w.r.t the outliers. Our main findings suggest that outliers re-identification via linkage attack is feasible and easily achieved. Furthermore, additional safeguards such as differential privacy can prevent re-identification, albeit at the expense of the data utility.