Membership Inference Attack

An attack on a machine-learning model that determines whether a particular record was part of the training data. Especially concerning for models trained on sensitive datasets (health, legal, financial).