A Noise-Resilient Semi-Supervised Graph Autoencoder for Overlapping Semantic Community Detection

A Noise-Resilient Semi-Supervised Graph Autoencoder for Overlapping Semantic Community Detection Abdelfateh Bekkair, Slimane Bellaouar and Slimane Oulad-Naoui Laboratoire des Mathématiques et Sciences Appliquées (LMSA), Université de Ghardaia, Ghardaia, Algeria Faculty of Sciences and Technology, Université de Ghardaia, Ghardaia, AlgeriaA R T I C L E I N F OKeywords: Overlapping community detection Graph attention autoencoder Semi-supervised learning Attributed networks Attribute noise analysis A B S T R A C T Community detection in networks with overlapping structures remains a significant challenge, particularly in noisy real-world environments where integrating topology, node attributes, and prior information is critical. To address this, we propose a semi-supervised graph autoencoder that combines graph multi-head attention and modularity maximization to robustly detect overlapping communities. The model learns semantic representations by fusing structural, attribute, and prior knowledge while explicitly addressing noise in node features. Key innovations include a noise-resistant architecture and a semantic semi-supervised design optimized for community quality through modularity constraints. Experiments demonstrate superior performance the model outperforms state-of-the-art methods in overlapping community detection (improvements in NMI and F1-score) and exhibits exceptional robustness to attribute noise, maintaining stable performance under 60% feature corruption.