In plain words: This survey gathers the proposed ways to score how private a synthetic table is, since no agreed rule exists yet. It lays out the options so researchers can choose and report privacy checks consistently.

Abstract · Privacy Measurement in Tabular Synthetic Data: State of the Art and Future Research Directions

Synthetic data (SD) have garnered attention as a privacy enhancing technology. Unfortunately, there is no standard for quantifying their degree of privacy protection. In this paper, we discuss proposed quantification approaches. This contributes to the development of SD privacy standards; stimulates multi-disciplinary discussion; and helps SD researchers make informed modeling and evaluation decisions.

Alexander Boudewijn, Andrea Filippo Ferraris, Daniele Panfilo, Vanessa Cocca, Sabrina Zinutti, Karel De Schepper, Carlo Rossi Chauvenet
arXiv:2311.17453 · cs.AI, cs.CR, cs.DB, stat.ML · submitted Nov 29, 2023
abstract · pdf · html · 20 pages, 4 tables, 8 figures; NeurIPS 2023 Workshop on Synthetic Data Generation with Generative AI