Google Research

On Avoiding the Union Bound When Answering Multiple Differentially Private Queries

Annual Conference on Learning Theory (COLT) (2021), pp. 2133-2146

Abstract

In this work, we study the problem of answering k queries with (ε, δ)-differential privacy, where each query has sensitivity one. We give a mechanism for this task that achieves an error bound of O(sqrt(k ln(1/δ))/ε), which is known to be tight (Steinke and Ullman, 2016).

A parallel work by Dagan and Kur (2020) provides a similar result, albeit via a completely different approach. One difference between our work and theirs is that our guarantee holds even when δ < 2^−Ω(k/(log k)^8) whereas theirs does not apply in this case. On the other hand, the algorithm of Dagan and Kur has a remarkable advantage that the ℓ∞ error bound of O(sqrt(k ln(1/δ))/ε) holds not only in expectation but always (i.e., with probability one) while we can only get a high probability (or expected) guarantee on the error.

Learn more about how we do research

We maintain a portfolio of research projects, providing individuals and teams the freedom to emphasize specific types of work