tapas.threat_models.aia

Threat Models for Attribute Inference Attacks.

Attribute Inference Attacks aim at inferring the value of a sensitive attribute for a target user, given some known attributes and the synthetic data.

Classes

AIALabeller(attacker_knowledge, ...[, ...])	Replace a record in the private dataset with a given target record, and randomly set the value of a given sensitive attribute in that record.
TargetedAIA(attacker_knowledge_data, ...[, ...])	This threat model implements a MIA with arbitrary attacker knowledge on data and generator.