Towards a More Reliable Privacy-preserving Recommender System

TL;DR

This paper introduces SDCF, a distributed recommendation framework that ensures comprehensive privacy preservation—including ratings, model, and existence—using differential privacy guarantees, with minimal accuracy loss demonstrated through experiments.

Contribution

The paper presents a novel distributed recommendation framework, SDCF, that guarantees differential privacy for ratings, model, and existence, enhancing privacy in collaborative filtering.

Findings

SDCF maintains privacy of ratings, model, and existence.

Experiments show minimal accuracy loss with privacy guarantees.

SDCF is applicable to rating prediction and action prediction tasks.

Abstract

This paper proposes a privacy-preserving distributed recommendation framework, Secure Distributed Collaborative Filtering (SDCF), to preserve the privacy of value, model and existence altogether. That says, not only the ratings from the users to the items, but also the existence of the ratings as well as the learned recommendation model are kept private in our framework. Our solution relies on a distributed client-server architecture and a two-stage Randomized Response algorithm, along with an implementation on the popular recommendation model, Matrix Factorization (MF). We further prove SDCF to meet the guarantee of Differential Privacy so that clients are allowed to specify arbitrary privacy levels. Experiments conducted on numerical rating prediction and one-class rating action prediction exhibit that SDCF does not sacrifice too much accuracy for privacy.