Differential Privacy
A data release mechanism is differentially private if any computation performed on the released data yields essentially the same result if a single data point is added/removed.
More specifically, let \(Q\) be any (probabilistic) query function. \(Q\) is \(\epsilon\)-differentially private if for all databases \(B\), \(B'\) that differ in one item, and for all functions \(F\), and all values \(y\):
$$ \mathbb{P}\{F(Q(B)) = y\} \le e^{\epsilon} \cdot \mathbb{P}\{F(Q(B')) = y\} $$
\(\epsilon\) is a privacy parameter. Increasing \(\epsilon\) improves accuracy, but reduces privacy.
See Differential Privacy Blog Series | NIST for more discussions on Differential Privacy.
Cynthia Dwork, the pioneer of Differential Privacy, has just won the 2020 Knuth Prize .