Privacy-preserving set union

Abstract

Recently there has been a significant amount of work on privacy-preserving set operations, including: set intersection [14,6,21,9], testing set disjointness [17], multi-set operations [18], and set union [16,1,18]. In this paper, we introduce novel protocols for privacy-preserving set union in the malicious adversary model. More specifically, each participant inputs a set of values, and at the end of the protocol, each participant learns the items that are in at least one participant's set without learning the frequency of the items or which participant(s) contributed specific items. To our knowledge our protocol is the most efficient privacy-preserving set union protocol for the malicious adversary model to date. © Springer-Verlag Berlin Heidelberg 2007.