Publication | Open Access
Graph Transformer Networks
514
Citations
0
References
2019
Year
Graph Neural NetworksNetwork ScienceGraph TheoryData ScienceGraph Representation LearningNetwork EstimationMachine LearningGraph Transformer NetworksEngineeringGraph Neural NetworkNetwork AnalysisGraph Signal ProcessingComputer ScienceHomogeneous GraphsGraph AnalysisDeep LearningGraph ProcessingRepresentation Learning
Graph neural networks have achieved state‑of‑the‑art performance on node classification and link prediction, yet they are primarily designed for fixed, homogeneous graphs and struggle with misspecified or heterogeneous graphs. This work introduces Graph Transformer Networks that generate new graph structures by identifying useful connections between previously unconnected nodes while learning node representations end‑to‑end. The core Graph Transformer layer learns a soft selection of edge types and composite relations to construct multi‑hop meta‑paths, enabling the creation of useful graph structures. Experiments demonstrate that GTNs automatically learn task‑specific graph structures and produce powerful node representations, achieving superior performance on three benchmark node‑classification tasks compared to methods that rely on domain‑specific meta‑paths.
Graph neural networks (GNNs) have been widely used in representation learning on graphs and achieved state-of-the-art performance in tasks such as node classification and link prediction. However, most existing GNNs are designed to learn node representations on the fixed and homogeneous graphs. The limitations especially become problematic when learning representations on a misspecified graph or a heterogeneous graph that consists of various types of nodes and edges. In this paper, we propose Graph Transformer Networks (GTNs) that are capable of generating new graph structures, which involve identifying useful connections between unconnected nodes on the original graph, while learning effective node representation on the new graphs in an end-to-end fashion. Graph Transformer layer, a core layer of GTNs, learns a soft selection of edge types and composite relations for generating useful multi-hop connections so-call meta-paths. Our experiments show that GTNs learn new graph structures, based on data and tasks without domain knowledge, and yield powerful node representation via convolution on the new graphs. Without domain-specific graph preprocessing, GTNs achieved the best performance in all three benchmark node classification tasks against the state-of-the-art methods that require pre-defined meta-paths from domain knowledge.