Multi-View Graph Feature Propagation for Privacy Preservation and Feature Sparsity
Etzion Harari, Moshe Unger
TL;DR
This work addresses node classification under extreme feature sparsity and privacy constraints by introducing Multi-view Feature Propagation (MFP), which extends classic Feature Propagation with multiple Gaussian-noised feature views. MFP uses stochastic sparse sampling to hide most attributes and then propagates several distinct masked views across the graph, concatenating the results to form a rich representation for a downstream GNN. The approach achieves strong predictive accuracy close to full-feature baselines while substantially reducing privacy leakage, as evidenced by RMSE/PCC analyses and cross-representation results that show the propagated outputs are not faithful reconstructions of the original features. Sensitivity analyses demonstrate robustness across homophily, number of views, and propagation depth, highlighting practical default settings for privacy-aware graph learning in real-world domains such as healthcare and finance.
Abstract
Graph Neural Networks (GNNs) have demonstrated remarkable success in node classification tasks over relational data, yet their effectiveness often depends on the availability of complete node features. In many real-world scenarios, however, feature matrices are highly sparse or contain sensitive information, leading to degraded performance and increased privacy risks. Furthermore, direct exposure of information can result in unintended data leakage, enabling adversaries to infer sensitive information. To address these challenges, we propose a novel Multi-view Feature Propagation (MFP) framework that enhances node classification under feature sparsity while promoting privacy preservation. MFP extends traditional Feature Propagation (FP) by dividing the available features into multiple Gaussian-noised views, each propagating information independently through the graph topology. The aggregated representations yield expressive and robust node embeddings. This framework is novel in two respects: it introduces a mechanism that improves robustness under extreme sparsity, and it provides a principled way to balance utility with privacy. Extensive experiments conducted on graph datasets demonstrate that MFP outperforms state-of-the-art baselines in node classification while substantially reducing privacy leakage. Moreover, our analysis demonstrates that propagated outputs serve as alternative imputations rather than reconstructions of the original features, preserving utility without compromising privacy. A comprehensive sensitivity analysis further confirms the stability and practical applicability of MFP across diverse scenarios. Overall, MFP provides an effective and privacy-aware framework for graph learning in domains characterized by missing or sensitive features.