Fast Differentially Private Matrix Factorization

Abstract

Differentially private collaborative filtering is a challenging task, both in terms of accuracy and speed. We present a simple algorithm that is provably differentially private, while offering good performance, using a novel connection of differential privacy to Bayesian posterior sampling via Stochastic Gradient Langevin Dynamics. Due to its simplicity the algorithm lends itself to efficient implementation. By careful systems design and by exploiting the power law behavior of the data to maximize CPU cache bandwidth we are able to generate 1024 dimensional models at a rate of 8.5 million recommendations per second on a single PC.

References in corpus (5)

• HOGWILD!: A Lock-Free Approach to Parallelizing Stochastic Gradient Descent
• GraphLab: A New Framework For Parallel Machine Learning
• Consistency and fluctuations for stochastic gradient Langevin dynamics
• Guess Who Rated This Movie: Identifying Users Through Subspace Clustering
• Large-Scale Distributed Bayesian Matrix Factorization using Stochastic Gradient MCMC

Cited by in corpus (4)

• CryptoRec: Privacy-preserving Recommendation as a Service
• PD-ML-Lite: Private Distributed Machine Learning from Lighweight Cryptography
• A Probabilistic View of Neighborhood-based Recommendation Methods
• Differentially Private Link Prediction With Protected Connections