Node-wise Localization of Graph Neural Networks

TL;DR

This paper introduces a node-wise localization method for GNNs that combines global and local information, improving node representations by adapting models to local contexts, and demonstrates superior performance on benchmark graphs.

Contribution

The paper proposes a novel node-wise localization approach for GNNs that balances global patterns with local node contexts, enhancing representation quality.

Findings

Consistently outperforms state-of-the-art GNNs on benchmark datasets.

Effectively captures local node context for improved representations.

Demonstrates the benefit of combining global and local modeling in GNNs.

Abstract

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.