TY - JOUR
T1 - Exploiting node-feature bipartite graph in graph convolutional networks
AU - Jiang, Yuli
AU - Lin, Huaijia
AU - Li, Ye
AU - Rong, Yu
AU - Cheng, Hong
AU - Huang, Xin
N1 - Funding Information (Section snippets):
The work was supported by grants from NSFC Grant No. U1936205, the Research Grants Council of the Hong Kong Special Administrative Region, China [Project No.: CUHK 14205618, CUHK 14217622, HKBU 22200320, HKBU 12200021], Tencent AI Lab RhinoBird Focused Research Program GF202101, and CUHK Direct Grant No. 4055159.
Publisher Copyright:
© 2023 Elsevier Inc.
PY - 2023/5
Y1 - 2023/5
N2 - In recent years, Graph Convolutional Networks (GCNs), which extend convolutional neural networks to graph structure, have achieved great success on many graph learning tasks by fusing structure and feature information, such as node classification. However, the graph structure is constructed from real-world data and usually contains noise or redundancy. In addition, this structural information is based on manually defined relations and is not potentially optimal for downstream tasks. In this paper, we utilize the knowledge from node features to enhance the expressive power of GCN models in a plug-and-play fashion. Specifically, we build a node-feature bipartite graph and exploit the bipartite graph convolutional network to model node-feature relations. By aligning results from the original graph structure and node-feature relations, we can make a more accurate prediction for each node in an end-to-end manner. Extensive experiments demonstrate that the proposed model can extract knowledge from two branches and improve the performance of various GCN models on typical graph data sets and 3D point cloud data.
AB - In recent years, Graph Convolutional Networks (GCNs), which extend convolutional neural networks to graph structure, have achieved great success on many graph learning tasks by fusing structure and feature information, such as node classification. However, the graph structure is constructed from real-world data and usually contains noise or redundancy. In addition, this structural information is based on manually defined relations and is not potentially optimal for downstream tasks. In this paper, we utilize the knowledge from node features to enhance the expressive power of GCN models in a plug-and-play fashion. Specifically, we build a node-feature bipartite graph and exploit the bipartite graph convolutional network to model node-feature relations. By aligning results from the original graph structure and node-feature relations, we can make a more accurate prediction for each node in an end-to-end manner. Extensive experiments demonstrate that the proposed model can extract knowledge from two branches and improve the performance of various GCN models on typical graph data sets and 3D point cloud data.
KW - Bipartite graph
KW - Bipartite graph convolutional networks
KW - Graph convolutional networks
KW - Node classification
KW - Semi-supervised learning
UR - http://www.scopus.com/inward/record.url?scp=85149805967&partnerID=8YFLogxK
U2 - 10.1016/j.ins.2023.01.107
DO - 10.1016/j.ins.2023.01.107
M3 - Journal article
AN - SCOPUS:85149805967
SN - 0020-0255
VL - 628
SP - 409
EP - 423
JO - Information Sciences
JF - Information Sciences
ER -