Empirically evaluating privacy in machine learning III: functional & gaussian differential privacy
The third in a series of blog posts on machine learning and privacy: why functional differential privacy is useful for auditing privacy in machine learning.