SynHING: Synthetic Heterogeneous Information Network Generation for Graph Learning and Explanation
This addresses the need for better evaluation tools for GNN explainers in complex networks, though it is incremental as it builds on existing synthetic graph generation methods.
The paper tackles the lack of robust baselines and datasets for evaluating GNN explainers in Heterogeneous Information Networks (HINs) by introducing SynHING, a framework for generating synthetic HINs that closely mimic original graphs' structural and statistical properties, providing ground-truth motifs for evaluation.
Graph Neural Networks (GNNs) excel in delineating graph structures in diverse domains, including community analysis and recommendation systems. As the interpretation of GNNs becomes increasingly important, the demand for robust baselines and expansive graph datasets is accentuated, particularly in the context of Heterogeneous Information Networks (HIN). Addressing this, we introduce SynHING, a novel framework for Synthetic Heterogeneous Information Network Generation aimed at enhancing graph learning and explanation. SynHING systematically identifies major motifs in a target HIN and employs a bottom-up generation process with intra-cluster and inter-cluster merge modules. This process, supplemented by post-pruning techniques, ensures the synthetic HIN closely mirrors the original graph's structural and statistical properties. Crucially, SynHING provides ground-truth motifs for evaluating GNN explainer models, setting a new standard for explainable, synthetic HIN generation and contributing to the advancement of interpretable machine learning in complex networks.