Recurrent Private Set Intersection for Unbalanced Databases with Cuckoo Hashing and Leveled FHE
Eduardo Chielle
Network and Distributed System Security (NDSS) Symposium 2025 · Day 3 · Privacy Preservation
This talk by Eduardo Chielle introduces a novel protocol for **Private Set Intersection (PSI)**, specifically designed to address the challenges of **recurrent** intersections involving **unbalanced databases**. PSI is a cryptographic primitive that allows two or more parties to compute the intersection of their private datasets without revealing any information about the elements not in the intersection. While traditional PSI protocols often assume sets of similar sizes and one-time computations, many real-world applications, such as private contact discovery or malicious URL detection, involve a small, frequently changing set intersecting with a large, static database.
AI review
Legitimate cryptographic research with a concrete, measurable contribution: a recurrent unbalanced PSI protocol that compresses encrypted sender state from 7 TB to 12.5 MB and cuts intersection latency to tens of milliseconds by combining Cuckoo Hashing with a novel CRT-based FHE packing layer. The work is technically honest, the gains are quantified against real baselines, and the implementation is public — this is not a paper that hides behind asymptotic claims.