Hop-Hop Relation-aware Graph Neural Networks
This work addresses the challenge of unifying GNN representation learning for both homogeneous and heterogeneous graphs, which is a significant problem for researchers and practitioners working with diverse graph structures.
This paper introduces Hop-Hop Relation-aware Graph Neural Network (HHR-GNN), a model designed to unify representation learning across both homogeneous and heterogeneous graphs. The model achieves competitive performance on five benchmarks, notably demonstrating a speedup of up to 13,000 times per training epoch on large heterogeneous graphs.
Graph Neural Networks (GNNs) are widely used in graph representation learning. However, most GNN methods are designed for either homogeneous or heterogeneous graphs. In this paper, we propose a new model, Hop-Hop Relation-aware Graph Neural Network (HHR-GNN), to unify representation learning for these two types of graphs. HHR-GNN learns a personalized receptive field for each node by leveraging knowledge graph embedding to learn relation scores between the central node's representations at different hops. In neighborhood aggregation, our model simultaneously allows for hop-aware projection and aggregation. This mechanism enables the central node to learn a hop-wise neighborhood mixing that can be applied to both homogeneous and heterogeneous graphs. Experimental results on five benchmarks show the competitive performance of our model compared to state-of-the-art GNNs, e.g., up to 13K faster in terms of time cost per training epoch on large heterogeneous graphs.