LSGNN: Towards General Graph Neural Network in Node Classification by Local Similarity
This addresses the problem of heterophily in node classification for GNN users, offering a plug-and-play module that can boost existing models, though it is incremental as it builds on prior work on multi-hop neighbor fusion.
The paper tackles the issue of heterophily in Graph Neural Networks (GNNs) by proposing LSGNN, which uses local similarity for node-level weighted fusion and an Initial Residual Difference Connection to improve multi-hop information extraction, achieving comparable or superior state-of-the-art performance on benchmark datasets.
Heterophily has been considered as an issue that hurts the performance of Graph Neural Networks (GNNs). To address this issue, some existing work uses a graph-level weighted fusion of the information of multi-hop neighbors to include more nodes with homophily. However, the heterophily might differ among nodes, which requires to consider the local topology. Motivated by it, we propose to use the local similarity (LocalSim) to learn node-level weighted fusion, which can also serve as a plug-and-play module. For better fusion, we propose a novel and efficient Initial Residual Difference Connection (IRDC) to extract more informative multi-hop information. Moreover, we provide theoretical analysis on the effectiveness of LocalSim representing node homophily on synthetic graphs. Extensive evaluations over real benchmark datasets show that our proposed method, namely Local Similarity Graph Neural Network (LSGNN), can offer comparable or superior state-of-the-art performance on both homophilic and heterophilic graphs. Meanwhile, the plug-and-play model can significantly boost the performance of existing GNNs. Our code is provided at https://github.com/draym28/LSGNN.