An Effective and Differentially Private Protocol for Secure Distributed Cardinality Estimation

Abstract

Counting the number of distinct elements distributed over multiple data holders is a fundamental problem with many real-world applications ranging from crowd counting to network monitoring. Although a number of space and computationally efficient sketch methods (e.g., the Flajolet-Martin sketch and the HyperLogLog sketch) for cardinality estimation have been proposed to solve the above problem, these sketch methods are insecure when considering privacy concerns related to the use of each data holder's personal dataset. Despite a recently proposed protocol that successfully implements the well-known Flajolet-Martin (FM) sketch on a secret-sharing based multiparty computation (MPC) framework for solving the problem of private distributed cardinality estimation (PDCE), we observe that this MPC-FM protocol is not differentially private. In addition, the MPC-FM protocol is computationally expensive, which limits its applications to data holders with limited computation resources. To address the above issues, in this paper we propose a novel protocol DP-DICE, which is computationally efficient and differentially private for solving the problem of PDCE. Experimental results show that our DP-DICE achieves orders of magnitude speedup and reduces the estimation error by several times in comparison with state-of-the-arts under the same security requirements.

Secure federated Boolean count queries using fully-homomorphic cryptography

“…We begin by giving our implementation of secure union cardinality. Although existing work like MPC-FM [13] and DP-DICE [19] also solve this problem, that work was concurrent to ours, and ours is both easy to implement and fairly efficient.…”

“…While suitable for complex analyses, such runtimes exceed reasonable expectations in the workflow of clinical query systems, such as i2b2 or Shrine, which are expected to return queries on the order of only a few minutes at most. The recent MPC-FM [13] and DP-DICE [19] algorithms fully solve what they call the secure distributed cardinality estimation problem, the latter on the order of minutes, which is very nearly fast enough. However, the secure distributed cardinality estimation task only resolves duplicate records across data custodians, and does not allow for combining partial information about the same patient, so those do not solve the federated Boolean count query problem we pose-in the notation above, these methods only solve the union half of the problem, but not the intersection portion.…”