<ul><li>Privacy vulnerabilities have been identified in marginals-based synthetic data generation.</li><li>Marginals-based synthetic data generation algorithms leak information about individuals that can be recovered more efficiently than previously understood.</li><li>A membership inference attack, MAMA-MIA, has been developed to exploit these vulnerabilities.</li><li>The attack allows for more accurate and faster learning of hidden data compared to other leading attacks.</li></ul>

Privacy Vulnerabilities in Marginals-based Synthetic Data

Discover more