1 paper
Bargav Jayaraman, David Evans
Models can expose sensitive information about their training data. In an attribute inference attack, an adversary has partial knowledge of some training records and access to a mod…