SoK: Differential Privacy on Graph-Structured Data
This work addresses privacy challenges for researchers and practitioners working with graph data, but it is incremental as it reviews and organizes existing formulations rather than introducing new methods.
The paper systematizes differential privacy formulations for graph-structured data, addressing challenges in privacy loss computation due to data interconnectivity and the absence of a standard DP formulation, and compares works on graph analysis and learning tasks including GNNs.
In this work, we study the applications of differential privacy (DP) in the context of graph-structured data. We discuss the formulations of DP applicable to the publication of graphs and their associated statistics as well as machine learning on graph-based data, including graph neural networks (GNNs). The formulation of DP in the context of graph-structured data is difficult, as individual data points are interconnected (often non-linearly or sparsely). This connectivity complicates the computation of individual privacy loss in differentially private learning. The problem is exacerbated by an absence of a single, well-established formulation of DP in graph settings. This issue extends to the domain of GNNs, rendering private machine learning on graph-structured data a challenging task. A lack of prior systematisation work motivated us to study graph-based learning from a privacy perspective. In this work, we systematise different formulations of DP on graphs, discuss challenges and promising applications, including the GNN domain. We compare and separate works into graph analysis tasks and graph learning tasks with GNNs. Finally, we conclude our work with a discussion of open questions and potential directions for further research in this area.