Differentially Private Decoupled Graph Convolutions for Multigranular Topology Protection

TL;DR

This paper introduces a novel framework called Graph Differential Privacy (GDP) and a new type of graph convolution, DPDGCs, to enhance privacy protection in graph neural networks while maintaining utility, addressing limitations of existing DP-GNNs.

Contribution

The paper proposes GDP for multigranular privacy in GNNs and introduces DPDGCs, a new convolution method that improves privacy-utility trade-offs over existing DP-GNN approaches.

Findings

DPDGCs outperform existing DP-GNNs in privacy-utility trade-offs.

Graph convolutions require noise scaling with maximum node degree, limiting privacy.

GDP provides flexible privacy guarantees for graph topology and node attributes.

Abstract

GNNs can inadvertently expose sensitive user information and interactions through their model predictions. To address these privacy concerns, Differential Privacy (DP) protocols are employed to control the trade-off between provable privacy protection and model utility. Applying standard DP approaches to GNNs directly is not advisable due to two main reasons. First, the prediction of node labels, which relies on neighboring node attributes through graph convolutions, can lead to privacy leakage. Second, in practical applications, the privacy requirements for node attributes and graph topology may differ. In the latter setting, existing DP-GNN models fail to provide multigranular trade-offs between graph topology privacy, node attribute privacy, and GNN utility. To address both limitations, we propose a new framework termed Graph Differential Privacy (GDP), specifically tailored to graph learning. GDP ensures both provably private model parameters as well as private predictions. Additionally, we describe a novel unified notion of graph dataset adjacency to analyze the properties of GDP for different levels of graph topology privacy. Our findings reveal that DP-GNNs, which rely on graph convolutions, not only fail to meet the requirements for multigranular graph topology privacy but also necessitate the injection of DP noise that scales at least linearly with the maximum node degree. In contrast, our proposed Differentially Private Decoupled Graph Convolutions (DPDGCs) represent a more flexible and efficient alternative to graph convolutions that still provides the necessary guarantees of GDP. To validate our approach, we conducted extensive experiments on seven node classification benchmarking and illustrative synthetic datasets. The results demonstrate that DPDGCs significantly outperform existing DP-GNNs in terms of privacy-utility trade-offs.