Naturally Private Recommendations with Determinantal Point Processes

TL;DR

This paper explores how Determinantal Point Processes (DPPs) can be adapted to satisfy differential privacy constraints, balancing recommendation diversity and privacy with minimal alterations.

Contribution

It introduces DPPs for privacy-preserving recommendations, derives privacy guarantees, analyzes their sensitivity, and proposes more efficient privacy-utility trade-offs.

Findings

DPPs can be modified to satisfy epsilon-Differential Privacy.

Sensitivity analysis of DPPs provides insights into privacy guarantees.

Proposed alternatives improve privacy-utility balance.

Abstract

Often we consider machine learning models or statistical analysis methods which we endeavour to alter, by introducing a randomized mechanism, to make the model conform to a differential privacy constraint. However, certain models can often be implicitly differentially private or require significantly fewer alterations. In this work, we discuss Determinantal Point Processes (DPPs) which are dispersion models that balance recommendations based on both the popularity and the diversity of the content. We introduce DPPs, derive and discuss the alternations required for them to satisfy epsilon-Differential Privacy and provide an analysis of their sensitivity. We conclude by proposing simple alternatives to DPPs which would make them more efficient with respect to their privacy-utility trade-off.