Paper 2025/1328

Private Set Intersection and other Set Operations in the Third Party Setting

Foo Yee Yeo
Jason H. M. Ying
Abstract

We present a collection of protocols to perform privacy-preserving set operations in the third-party private set intersection (PSI) setting. This includes several protocols for multi-party third party PSI. In this model, there are multiple input parties (or clients) each holding a private set of elements and the receiver is an external party (termed as third-party) with no inputs. Multi-party third party PSI enables the receiver to learn only the intersection result of all input clients' private sets while revealing nothing else to the clients and the receiver. Our solutions include constructions that are provably secure against an arbitrary number of colluding parties in the semi-honest model. Additionally, we present protocols for third-party private set difference and private symmetric difference, whereby the learned output by the inputless third-party is the set difference and symmetric difference respectively of two other input parties, while preserving the same privacy guarantees. The motivation in the design of these protocols stems from their utilities in numerous real-world applications. We implemented our protocols and conducted experiments across various input and output set sizes.

Metadata
Available format(s)
PDF
Category
Cryptographic protocols
Publication info
Published elsewhere. 34th USENIX Security Symposium (USENIX Security '25)
Keywords
private set operationsprivate set intersectionPSI
Contact author(s)
fooyee yeo @ seagate com
jasonhweiming ying @ seagate com
History
2025-07-22: approved
2025-07-21: received
See all versions
Short URL
https://ia.cr/2025/1328
License
Creative Commons Attribution-NonCommercial-NoDerivs
CC BY-NC-ND

BibTeX

@misc{cryptoeprint:2025/1328,
      author = {Foo Yee Yeo and Jason H. M. Ying},
      title = {Private Set Intersection and other Set Operations in the Third Party Setting},
      howpublished = {Cryptology {ePrint} Archive, Paper 2025/1328},
      year = {2025},
      url = {https://eprint.iacr.org/2025/1328}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.