Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I recently read the definition of differential privacy, which is as follows:

a randomized function $K$ gives $\epsilon$-differential privacy if for all data sets $D$ and $D'$ differing in at most one row, and all output space $S$, we have $Pr[K(D) \in S] \leq \exp(\epsilon).Pr[K(D') \in S]$ where the probability space in each case is over the coin flips of $K$

Could somebody explain what it means when saying "probability is over the coin flip" ? an example would be greatly appreciated

Cheers

share|improve this question

1 Answer 1

the probability space in each case is over the coin flips of K

Simply put, $K$ has values in a discrete set $\mathfrak K$ with $\Pr(K=k)=p_k$ for every $k$ in $\mathfrak K$, and $$ \Pr(K(D)\in S)=\sum_{k\in\mathfrak K}p_k\cdot\mathbf 1_{k(D)\in S}. $$ Similarly for $\Pr(K(D')\in S)$.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.