Predicting Privacy Leakage from Weight Spectral Density

Richard J. Preen, Jim Smith

Membership inference attacks (MIAs) are widely used to audit the privacy disclosure risk of machine learning models, however current state-of-the-art attacks require training compu…