Pure Differentially Private Summation from Anonymous Messages

The shuffled (aka anonymous) model has recently generated significant interest as a candidate distributed privacy framework with trust assumptions better than the central model but with achievable error rates smaller than the local model. In this paper, we study pure differentially private protocols in the shuffled model for summation, a very basic and widely used primitive. Specifically: - For the binary summation problem where each of n users holds a bit as an input, we give a pure ε-differentially private protocol for estimating the number of ones held by the users up to an absolute error of O⏔ (1), and where each user sends O⏔ (log n) one-bit messages. This is the first pure protocol in the shuffled model with error o (√n) for constant values of ε. our binary summation protocol as a building block, we give a pure ε-differentially private protocol that performs summation of real numbers in 0, 1 up to an absolute error of O⏔ (1), and where each user sends O⏔ (log³ n) messages each consisting of O (log log n) bits. - In contrast, we show that for any pure ε-differentially private protocol for binary summation in the shuffled model having absolute error n^0. 5-Ω (1), the per user communication has to be at least Ω⏔ (√log n) bits. This implies (i) the first separation between the (bounded-communication) multi-message shuffled model and the central model, and (ii) the first separation between pure and approximate differentially private protocols in the shuffled model. Interestingly, over the course of proving our lower bound, we have to consider (a generalization of) the following question that might be of independent interest: given γ ∈ (0, 1), what is the smallest positive integer m for which there exist two random variables X⁰ and X¹ supported on 0, …, m such that (i) the total variation distance between X⁰ and X¹ is at least 1 - γ, and (ii) the moment generating functions of X⁰ and X¹ are within a constant factor of each other everywhere? We show that the answer to this question is m = Θ (√log (1/γ) ).