TAMIS: Tailored Membership Inference Attacks on Synthetic Data

Abstract

Membership Inference Attacks (MIA) enable to empirically assess the privacy of a machine learning algorithm. In this paper, we propose TAMIS, a novel MIA against differentially-private synthetic data generation methods that rely on graphical models. This attack builds upon MAMA-MIA, a recently-published state-of-the-art method. It lowers its computational cost and requires less attacker knowledge. Our attack is the product of a two-fold improvement. First, we recover the graphical model having generated a synthetic dataset by using solely that dataset, rather than shadow-modeling over an auxiliary one. This proves less costly and more performant. Second, we introduce a more mathematically-grounded attack score, that provides a natural threshold for binary predictions. In our experiments, TAMIS achieves better or similar performance as MAMA-MIA on replicas of the SNAKE challenge.

Cite

Text

Andrey et al. "TAMIS: Tailored Membership Inference Attacks on Synthetic Data." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2025. doi:10.1007/978-3-032-06096-9_12

Markdown

[Andrey et al. "TAMIS: Tailored Membership Inference Attacks on Synthetic Data." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2025.](https://mlanthology.org/ecmlpkdd/2025/andrey2025ecmlpkdd-tamis/) doi:10.1007/978-3-032-06096-9_12

BibTeX

@inproceedings{andrey2025ecmlpkdd-tamis,
  title     = {{TAMIS: Tailored Membership Inference Attacks on Synthetic Data}},
  author    = {Andrey, Paul and Le Bars, Batiste and Tommasi, Marc},
  booktitle = {European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases},
  year      = {2025},
  pages     = {203-220},
  doi       = {10.1007/978-3-032-06096-9_12},
  url       = {https://mlanthology.org/ecmlpkdd/2025/andrey2025ecmlpkdd-tamis/}
}