Counting Distinct Elements Under Person-Level Differential Privacy

Published: 21 Sept 2023, Last Modified: 02 Nov 2023NeurIPS 2023 posterEveryoneRevisionsBibTeX
Keywords: differential privacy, user-level privacy, person-level privacy, sensitivity
TL;DR: We show how to privately compute the number of unique elements, despite this having infinite sensitivity in the user-level DP setting.
Abstract: We study the problem of counting the number of distinct elements in a dataset subject to the constraint of differential privacy. We consider the challenging setting of person-level DP (a.k.a. user-level DP) where each person may contribute an unbounded number of items and hence the sensitivity is unbounded. Our approach is to compute a bounded-sensitivity version of this query, which reduces to solving a max-flow problem. The sensitivity bound is optimized to balance the noise we must add to privatize the answer against the error of the approximation of the bounded-sensitivity query to the true number of unique elements.
Supplementary Material: pdf
Submission Number: 24