Locally Private Graph Neural Networks

Graph Neural Networks (GNNs) have demonstrated superior performance in learning node representations for various graph inference tasks. However, learning over graph data can raise privacy concerns when nodes represent people or human-related variables that involve sensitive or personal information. While numerous techniques have been proposed for privacy-preserving deep learning over non-relational data, such as image, audio, video, and text, there is less work addressing the privacy issues pertained to applying deep learning algorithms on graphs. As a result and for the first time, in this paper, we study the problem of node-level privacy, where graph nodes have potentially sensitive features that need to be kept private, but they could be beneficial for a central server for training a GNN over the graph. To address this problem, we develop a privacy-preserving GNN learning algorithm with formal privacy guarantees based on Local Differential Privacy (LDP). Specifically, we propose an optimized LDP encoder and an unbiased rectifier, using which the server can communicate with the graph nodes to privately collect their data and approximate the graph convolution layer of the GNN. To further reduce the effect of the injected noise, we propose a simple graph convolution layer based on the multi-hop aggregation of the nodes' features. We argue why LDP is a better choice to tackle this problem compared to other privacy-preserving learning paradigms, such as federated learning, and discuss how GNNs, due to their unique internal structure, can be more robust to differentially private input perturbations than other deep learning models. Extensive experiments conducted over real-world datasets demonstrate the significant capability of our method in maintaining an appropriate privacy-accuracy trade-off.