Space4HGNN: A Novel, Modularized and Reproducible Platform to Evaluate Heterogeneous Graph Neural Network

Abstract

Heterogeneous Graph Neural Network (HGNN) has been successfully employed in various tasks, but we cannot accurately know the importance of different design dimensions of HGNNs due to diverse architectures and applied scenarios. Besides, in the research community of HGNNs, implementing and evaluating various tasks still need much human effort. To mitigate these issues, we first propose a unified framework covering most HGNNs, consisting of three components: heterogeneous linear transformation, heterogeneous graph transformation, and heterogeneous message passing layer. Then we build a platform Space4HGNN by defining a design space for HGNNs based on the unified framework, which offers modularized components, reproducible implementations, and standardized evaluation for HGNNs. Finally, we conduct experiments to analyze the effect of different designs. With the insights found, we distill a condensed design space and verify its effectiveness.

The Qilihe Street Pedestrian Bridge in Nanjing, China

“…Meta-path plays an essential role in the existing heterogeneous GNNs. However, recent researchers [17] experimentally found that homogeneous GNNs such as GAT [18] actually perform pretty well on heterogeneous graph by revisiting the model design, data preprocessing, and experimental settings of the heterogeneous GNNs, which calls into question the necessity of meta-paths [17], [19]. To answer this question, we first give an in-depth analysis about the intrinsic difference about meta-path-based models (e.g., HAN, MAGNN, GTN) and meta-path-free models (e.g., RGCN [20], GAT, Simple-HGN [17]).…”