Efficient Multi-Party Private Set Intersection Against Malicious Adversaries

Private set intersection (PSI) enables parties to compute the intersection of their inputs without leaking any additional information. Recently, there have been significant advances in the two-party settings with malicious security, making two-party PSI truly practical even compared with the naive insecure method. However, efficient PSI solutions in the more general case (multi-party) are only known for the semi-honest setting. How to construct a practical multiparty solution (especially for large inputs) in the malicious setting remains an important open question in this research line. In this work, we first present an efficient multi-party PSI protocol secure against any malicious adversary that does not simultaneously corrupt two designated parties. Next, we describe how to efficiently enhance security of the protocol when the two designated parties are corrupted. Finally, we present our experiment results for the first protocol, confirming that the efficiency scales linearly in the number of parties of the work Rindal and Rosulek (Eurocrypt 17).