Apple Inc.
PRACTICAL PRIVATE ALGORITHMS FOR ROBUST STATISTICS

Abstract:

Embodiments described herein provide a privacy mechanism to protect user data when transmitting the data to a server that estimates a p-th frequency moment, F.sub.p for p.di-elect cons.[1, 2] and l.sub.p low-rank approximation for p.di-elect cons.[1, 2). The privacy mechanism uses an encode-shuffle then analyze (ESA) framework that provides a compromise between the central and local model of privacy.