北京雁栖湖应用数学研究院

We will focus on the studies of special variants of Private Set Intersection, Private Set Union, and Multi-party Private Set Union, which allow two or more parties to compute set intersections or unions without revealing any additional information.

PSI has found numerous applications including privacy-preserving location sharing, private contact discovery, DNA testing and pattern matching. Similarly, PSU also has many applications in practice, such as cyber risk assessment and management, IP blacklist and vulnerability data aggregation, private DB supporting full join and private ID.

The most efficient PSI protocols [KKRT16, PRTY19, CM20, GPR+21, RS21] mainly rely on symmetric-key operations, except O(κ) public-key operations (where κ is a computational security parameter) in base Oblivious Transfer(OT) used in the OT extension protocol. There are also many variations of PSI, such as PSI-CA [IKN+20], Fuzzy PSI [BP24], Updatable PSI [BMS24] et al. Existing PSU protocols can be broadly divided into two categories based on the underlying cryptographic techniques used. The first category mainly relies on public-key techniques [KS05, Fri07, HN10, DC17], while the second category mainly relies on symmetric key techniques [KRTW19, GMR+21, JSZ+22]. Also depending on the usage scenario, the PSU protocols can be divided into balanced [CZZ+24,ZCL+23] and unbalanced [JSZ22,TCL+23,ZCL+24]. This course will track and incorporate the latest developments in the state-of-the-art research discussed above.