Privacy Preserving Aggregation of Ad Conversions using MPC

Learn about implementing privacy-preserving advertisement measurement systems through secure multi-party computation in this 34-minute conference talk. Explore how to privately join advertisement impressions and conversions while aggregating data across billions of users using a three-party honest majority MPC system. Discover the technical implementation details including secure evaluation of the DY PRF, shuffle operations, vectorized Boolean operations, and distributed zero-knowledge proofs. Examine the scale testing results demonstrating the system's capability to process 2 billion impressions and conversions across 3 parties with 50 shards each, providing practical insights into deploying privacy-preserving technologies for large-scale advertising analytics.