Contrastive Learning-Based privacy metrics in Tabular Synthetic Datasets

2025-02-19

AI Summary

Existing privacy evaluation methods for synthetic tabular data lack interpretable and quantifiable privacy guarantees. Method: This paper proposes an embedding-based privacy metric grounded in contrastive learning, which uniformly maps heterogeneous attributes into a measurable embedding space—enabling intuitive distance computation and modeling of privacy attacks such as membership inference. Contribution/Results: To our knowledge, this is the first work to apply contrastive learning to quantitative privacy assessment of synthetic data, effectively addressing the challenge of adapting to diverse attribute types (e.g., categorical, numerical, ordinal). The proposed lightweight metric requires no explicit modeling of GDPR-compliance constraints, yet achieves evaluation performance comparable to complex regulatory-compliance models across multiple public benchmarks. It offers significant advantages in computational efficiency and deployment simplicity while preserving rigorous privacy quantification.

Technology Category

• Machine Learning: Privacy
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Computer Vision: Bias, Fairness & Privacy

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

Synthetic data has garnered attention as a Privacy Enhancing Technology (PET) in sectors such as healthcare and finance. When using synthetic data in practical applications, it is important to provide protection guarantees. In the literature, two family of approaches are proposed for tabular data: on the one hand, Similarity-based methods aim at finding the level of similarity between training and synthetic data. Indeed, a privacy breach can occur if the generated data is consistently too similar or even identical to the train data. On the other hand, Attack-based methods conduce deliberate attacks on synthetic datasets. The success rates of these attacks reveal how secure the synthetic datasets are. In this paper, we introduce a contrastive method that improves privacy assessment of synthetic datasets by embedding the data in a more representative space. This overcomes obstacles surrounding the multitude of data types and attributes. It also makes the use of intuitive distance metrics possible for similarity measurements and as an attack vector. In a series of experiments with publicly available datasets, we compare the performances of similarity-based and attack-based methods, both with and without use of the contrastive learning-based embeddings. Our results show that relatively efficient, easy to implement privacy metrics can perform equally well as more advanced metrics explicitly modeling conditions for privacy referred to by the GDPR.

Problem

• Improves privacy assessment in synthetic datasets
• Overcomes data type and attribute obstacles
• Compares similarity-based and attack-based methods

Innovation

• Contrastive Learning improves privacy assessment
• Embeds data in more representative space
• Enables intuitive distance metrics for measurements