TAPAS: a Toolbox for Adversarial Privacy Auditing of Synthetic Data

2022/11/12 by Florimond Houssiau, J.B. Jordon, Houssiau, Florimond +15 · 9 citations
Computer Science · Decision Sciences · Social Sciences · #Artificial Intelligence (cs.AI) #Cryptography and Security (cs.CR) #Data Quality and Management #FOS: Computer and information sciences #Machine Learning (cs.LG) #Privacy, Security, and Data Protection #Privacy-Preserving Technologies in Data

paper · doi:10.48550/arxiv.2211.06550

openalex publication_date 2022/11/12 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/29

Abstract

Personal data collected at scale promises to improve decision-making and accelerate innovation. However, sharing and using such data raises serious privacy concerns. A promising solution is to produce synthetic data, artificial records to share instead of real data. Since synthetic records are not linked to real persons, this intuitively prevents classical re-identification attacks. However, this is insufficient to protect privacy. We here present TAPAS, a toolbox of attacks to evaluate synthetic data privacy under a wide range of scenarios. These attacks include generalizations of prior works and novel attacks. We also introduce a general framework for reasoning about privacy threats to synthetic data and showcase TAPAS on several examples.

Cited by

• It does what it says on the tin: safe synthetic data from coarsened margins
• When Tables Leak: Attacking String Memorization in LLM-Based Tabular Data Generation
• Beyond the Trade-off Curve: Multivariate and Advanced Risk-Utility Maps for Evaluating Anonymized and Synthetic Data
• SMOTE and Mirrors: Exposing Privacy Leakage from Synthetic Minority Oversampling
• Synth-MIA: A Testbed for Auditing Privacy Leakage in Tabular Data Synthesis
• Ensembling Membership Inference Attacks Against Tabular Generative Models
• Privacy Auditing Synthetic Data Release through Local Likelihood Attacks
• Privacy-Preserving Tabular Synthetic Data Generation Using TabularARGN
• Towards Useful and Private Synthetic Omics: Community Benchmarking of Generative Models for Transcriptomics Data

Related

• Phantoms and Disclosures: A Statistical Framework for Auditing Privacy in Synthetic Data
• Synthetic Data Outliers: Navigating Identity Disclosure
• Empirical Evaluation of Structured Synthetic Data Privacy Metrics: Novel experimental framework
• A Unified Framework for Quantifying Privacy Risk in Synthetic Data
• Synthetic Data: Methods, Use Cases, and Risks