Educational Resource / August 2020

“What? So What? Now What?”: Differential Privacy


image showing the trade off between accuracy and privacy as defined by epsilon
The Greek symbol ε, or epsilon, is used to define the privacy loss associated with the release of data from a data set.

Differential privacy has already gained widespread adoption by governments, firms, and researchers. It is already being used for “disclosure avoidance” by the U.S. census, for example, and Apple uses differential privacy to analyze user data ranging from emoji suggestions to Safari crashes. Google has even released an open-source version of a differential privacy library used in many of the company’s core products.

Read more on the CLTC Bulletin.