Heterogeneous Graph Transformer

Трансформер для неоднородных графов
Ziniu Hu, Yuxiao Dong, Kuansan Wang, Yizhou Sun
2020-04-20

HGSamplingHeterogeneous Graph TransformerOpen Academic Graphheterogeneous attentionheterogeneous graphs
Recent years have witnessed the emerging success of graph neural networks (GNNs) for modeling structured data. However, most GNNs are designed for homogeneous graphs, in which all nodes and edges belong to the same types, making it infeasible to represent heterogeneous structures. In this paper, we present the Heterogeneous Graph Transformer (HGT) architecture for modeling Web-scale heterogeneous graphs. To model heterogeneity, we design node- and edge-type dependent parameters to characterize the heterogeneous attention over each edge, empowering HGT to maintain dedicated representations for different types of nodes and edges. To handle Web-scale graph data, we design the heterogeneous mini-batch graph sampling algorithm—HGSampling—for efficient and scalable training. Extensive experiments on the Open Academic Graph of 179 million nodes and 2 billion edges show that the proposed HGT model consistently outperforms all the state-of-the-art GNN baselines by 9–21 on various downstream tasks. The dataset and source code of HGT are publicly available at https://github.com/acbull/pyHGT.
1
HGT maintains dedicated representations for different node and edge types, enabling more effective modeling of heterogeneous relational structures.
2
On the 179-million-node, 2-billion-edge Open Academic Graph, HGT consistently outperforms state-of-the-art GNN baselines by 9–21 across downstream tasks.
3
The HGSampling algorithm provides efficient, scalable heterogeneous mini-batch training for very large graph data.
4
The HGT dataset and source code are publicly released, supporting reproducibility and further research.
5
The Heterogeneous Graph Transformer (HGT) models Web-scale heterogeneous graphs using node- and edge-type-dependent parameters in its attention mechanism.

Web-scale heterogeneous graphs (e.g., Open Academic Graph with 179 million nodes and 2 billion edges)

Type-specific graph representation learning and scalable modeling performance on downstream tasks

Publication Details
Publication Date
2020-04-20
Journal
Publisher
ISSN
Cited by
2443
Access Type
Author Information
Authors
Ziniu Hu
Yuxiao Dong
Kuansan Wang
Yizhou Sun
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%