Beyond Low-Pass Filters: Adaptive Feature Propagation on Graphs
This addresses the generalizability problem for GNN users on diverse graph types, though it is incremental as it builds on existing spectral filter methods.
The paper tackled the limitation of graph neural networks (GNNs) being restricted to graphs with local homophily by proposing a flexible GNN model that handles both homophilic and heterophilic graphs, achieving state-of-the-art performance on heterophilic graphs and comparable results on homophilic ones across eight benchmark datasets.
Graph neural networks (GNNs) have been extensively studied for prediction tasks on graphs. As pointed out by recent studies, most GNNs assume local homophily, i.e., strong similarities in local neighborhoods. This assumption however limits the generalizability power of GNNs. To address this limitation, we propose a flexible GNN model, which is capable of handling any graphs without being restricted by their underlying homophily. At its core, this model adopts a node attention mechanism based on multiple learnable spectral filters; therefore, the aggregation scheme is learned adaptively for each graph in the spectral domain. We evaluated the proposed model on node classification tasks over eight benchmark datasets. The proposed model is shown to generalize well to both homophilic and heterophilic graphs. Further, it outperforms all state-of-the-art baselines on heterophilic graphs and performs comparably with them on homophilic graphs.