Ego-GNNs: Exploiting Ego Structures in Graph Neural Networks
This addresses a fundamental representational bottleneck in GNNs for graph-structured data, offering a novel method to improve performance in tasks like node classification.
The paper tackles the limitation of graph neural networks (GNNs) in recognizing triangles due to their tree-structured inductive bias, proposing Ego-GNNs that augment message-passing with ego graph information and showing they are provably more powerful, with experimental gains in node classification.
Graph neural networks (GNNs) have achieved remarkable success as a framework for deep learning on graph-structured data. However, GNNs are fundamentally limited by their tree-structured inductive bias: the WL-subtree kernel formulation bounds the representational capacity of GNNs, and polynomial-time GNNs are provably incapable of recognizing triangles in a graph. In this work, we propose to augment the GNN message-passing operations with information defined on ego graphs (i.e., the induced subgraph surrounding each node). We term these approaches Ego-GNNs and show that Ego-GNNs are provably more powerful than standard message-passing GNNs. In particular, we show that Ego-GNNs are capable of recognizing closed triangles, which is essential given the prominence of transitivity in real-world graphs. We also motivate our approach from the perspective of graph signal processing as a form of multiplex graph convolution. Experimental results on node classification using synthetic and real data highlight the achievable performance gains using this approach.