Simple and Efficient Heterogeneous Graph Neural Network

6 Jul 2022  ·  Xiaocheng Yang, Mingyu Yan, Shirui Pan, Xiaochun Ye, Dongrui Fan ·

Heterogeneous graph neural networks (HGNNs) have powerful capability to embed rich structural and semantic information of a heterogeneous graph into node representations. Existing HGNNs inherit many mechanisms from graph neural networks (GNNs) over homogeneous graphs, especially the attention mechanism and the multi-layer structure. These mechanisms bring excessive complexity, but seldom work studies whether they are really effective on heterogeneous graphs. This paper conducts an in-depth and detailed study of these mechanisms and proposes Simple and Efficient Heterogeneous Graph Neural Network (SeHGNN). To easily capture structural information, SeHGNN pre-computes the neighbor aggregation using a light-weight mean aggregator, which reduces complexity by removing overused neighbor attention and avoiding repeated neighbor aggregation in every training epoch. To better utilize semantic information, SeHGNN adopts the single-layer structure with long metapaths to extend the receptive field, as well as a transformer-based semantic fusion module to fuse features from different metapaths. As a result, SeHGNN exhibits the characteristics of simple network structure, high prediction accuracy, and fast training speed. Extensive experiments on five real-world heterogeneous graphs demonstrate the superiority of SeHGNN over the state-of-the-arts on both accuracy and training speed.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Heterogeneous Node Classification ACM (Heterogeneous Node Classification) SeHGNN Micro-F1 93.87 # 3
Claim Classification Macro-F1 93.95 # 1
Heterogeneous Node Classification DBLP (Heterogeneous Node Classification) SeHGNN Micro-F1 95.24 # 3
Macro-F1 94.86 # 1
Heterogeneous Node Classification Freebase (Heterogeneous Node Classification) SeHGNN Micro-F1 63.41 # 4
Macro-F1 50.71 # 1
Heterogeneous Node Classification IMDB (Heterogeneous Node Classification) SeHGNN Micro-F1 68.21 # 3
Macro-F1 66.63 # 1
Heterogeneous Node Classification OAG-L1-Field SeHGNN NDCG 86.01 # 3
MRR 84.95 # 3
Heterogeneous Node Classification OAG-Venue SeHGNN NDCG 46.75 # 5
MRR 29.11 # 5
Node Property Prediction ogbn-mag SeHGNN (ComplEx embs) Test Accuracy 0.5719 ± 0.0012 # 9
Validation Accuracy 0.5917 ± 0.0009 # 7
Number of params 8371231 # 16
Ext. data No # 1
Node Property Prediction ogbn-mag SeHGNN Test Accuracy 0.5671 ± 0.0014 # 10
Validation Accuracy 0.5870 ± 0.0008 # 10
Number of params 8371231 # 16
Ext. data No # 1

Methods