Private Set Intersection via Searchable Encryption against Malicious Adversaries

Shao et al. proposed a novel method for computing private set intersection based on searchable encryption. Compared with the traditional protocols, their technique requires much less computation for the client to obtain the intersection. Concretely, the client only computes m (the cardinality of the client's input set) multiplications, m hashes, and one modular exponentiation. However, their protocol only satisfies security against semi-honest behaviors. In this paper, we aim to promote the security of Shao et al.'s protocol. Our solution woks in the CRS model, and is based on the non-interactive zero knowledge proof and the succinct non-interactive argument. The proposed protocol satisfies security against malicious adversaries and needs only one round interaction.