Explore a conference talk from PEPR '23 that delves into protecting user privacy in Private Set Intersection (PSI) and the journey towards mitigating user tracking. Learn about the vulnerabilities of common PSIs that disclose intersection size, potentially revealing over 1% of one party's users to another in practical ad measurement scenarios. Discover the DPCA-PSI protocol, a novel approach that combines PSI with a two-party differentially private mechanism to securely compute intersection-related statistics while preserving privacy. Gain insights into applications of PSI in ad effectiveness measurement and cloud storage operations for detecting child exploitation material. Understand how this open-source solution provides researchers and practitioners with a secure means of working with private datasets while maintaining differential privacy protection.
Overview
Syllabus
PEPR '23 - Protecting User Privacy in Private Set Intersection: A Journey Toward Mitigating User
Taught by
USENIX