Node-wise Localization of Graph Neural Networks
This work addresses the challenge of local node variability in graph representation learning, offering a novel approach that could enhance GNN applications in domains like social networks or bioinformatics, though it appears incremental as it builds on existing GNN frameworks.
The paper tackles the problem of varying node distributions in graphs by proposing a node-wise localization method for Graph Neural Networks (GNNs) that adapts to local contexts, resulting in consistent performance improvements over state-of-the-art GNNs on four benchmark graphs.
Graph neural networks (GNNs) emerge as a powerful family of representation learning models on graphs. To derive node representations, they utilize a global model that recursively aggregates information from the neighboring nodes. However, different nodes reside at different parts of the graph in different local contexts, making their distributions vary across the graph. Ideally, how a node receives its neighborhood information should be a function of its local context, to diverge from the global GNN model shared by all nodes. To utilize node locality without overfitting, we propose a node-wise localization of GNNs by accounting for both global and local aspects of the graph. Globally, all nodes on the graph depend on an underlying global GNN to encode the general patterns across the graph; locally, each node is localized into a unique model as a function of the global model and its local context. Finally, we conduct extensive experiments on four benchmark graphs, and consistently obtain promising performance surpassing the state-of-the-art GNNs.