Differential privacy
What does Differential privacy mean?
A framework that bounds how much an output distribution can change when one person's data changes.
A practical example
Noise is calibrated to limit what a statistic reveals about an individual.