Differentially Private Uniformly Most Powerful Tests for Binomial Data

Furthermore, we obtain exactp-values, which are easily computed in terms of the Tulap random variable. We show that our results also apply to distribution-free hypothesis testsforcontinuous data.