Abstract:Most social networks can be modeled as heterogeneous graphs. Recently, advanced graph learning methods exploit the rich node properties and topological relationships for downstream tasks. That means that more private information is embedded in the representation. However, the existing privacy-preserving methods only focus on protecting the single type of node attributes or relationships, which neglect the significance of high-order semantic information. To address this issue, we propose a novel Heterogeneous graph neural network with Semantic-aware Differential privacy Guarantees named HeteSDG, which provides a double privacy guarantee and performance trade-off in terms of both graph features and topology. In particular, we first reveal the privacy leakage in heterogeneous graphs and define a membership inference attack with a semantic enhancement (MIS) that will improve the means of member inference attacks by obtaining side background knowledge through semantics. Then we design a two-stage mechanism, which includes the feature attention personalized mechanism and the topology gradient perturbation mechanism, where the privacy-preserving technologies are based on differential privacy. These mechanisms will defend against MIS and provide stronger interpretation, but simultaneously bring in noise for representation learning. To better balance the noise perturbation and learning performance, we utilize a bi-level optimization pattern to allocate a suitable privacy budget for the above two modules. Our experiments on four public benchmarks conduct performance experiments, ablation studies, inference attack verification, etc. The results show the privacy protection capability and generalization of HeteSDG.

Privacy Attack and Defense in Network Embedding

SAME: Sampling Attack in Multiplex Network Embedding.

Cross-Network Social User Embedding with Hybrid Differential Privacy Guarantees

Devil in Disguise: Breaching Graph Neural Networks Privacy through Infiltration

Privacy-Preserving Network Embedding Against Private Link Inference Attacks

Adversarial Privacy-Preserving Graph Embedding Against Inference Attack

Privacy-Preserving Graph Embedding based on Local Differential Privacy

Quantifying Privacy Leakage in Graph Embedding

Privacy Protection and Utility Trade-Off for Social Graph Embedding

Heterogeneous graph neural network with semantic-aware differential privacy guarantees

Secure Network Release with Link Privacy

Privacy Preservation in Social Network Against Public Neighborhood Attacks

Inference Attacks Against Graph Neural Networks

Unveiling Privacy Vulnerabilities: Investigating the Role of Structure in Graph Data

Anchor Link Prediction Via Network Structural Role for Privacy Leakage in Edge Computing

Enhancing Social Network Privacy with Accumulated Non-Zero Prior Knowledge

Anchor Link Prediction for Privacy Leakage Via De-Anonymization in Multiple Social Networks

IEA-DP: Information Entropy-driven Adaptive Differential Privacy Protection Scheme for social networks

Protecting sensitive relationships against inference attacks in social networks

Poincaré Differential Privacy for Hierarchy-Aware Graph Embedding

Seamless Privacy: Privacy-Preserving Subgraph Counting in Interactive Social Network Analysis