Inexact Unlearning Needs More Careful Evaluations to Avoid a False Sense of Privacy

Open in new window