It’s too noisy in here: using projection to improve Differential Privacy on RDF graphs
Résumé
In the last decade, adaptation of differential privacy to graph data has received growing attention. Most efforts have been dedicated to unlabeled homogeneous graphs, while labeled graphs with an underlying
semantic (e.g. RDF) have been mildly addressed.
In this paper, we present a new approach based on graph projection to adapt differential privacy to RDF graphs, while reducing query sensitivity. We propose three edge-addition based graph projection methods that transform the original RDF graph into a graph with bounded degree, bounded out-degree, and bounded typed-out-degree. We demonstrate that these projections preserve neighborhood, allowing to expand the domain of any differentially private algorithm from graphs with bounded (out/typed-out) degree to any arbitrary RDF graph. Experimental and analytical evaluation through a realistic twitter use-case shows that projection can provide two orders of magnitude of utility improvement.