Graph neural architecture search: A survey

• DOI: 10.26599/tst.2021.9010057
• Publication Date: 2021-12-09T21:09:15Z
• Authors: Babatounde Moctar...; Jianliang Gao; Jiamin Chen; Tengfei Lyu; Raeed Al-Sabri
• Keywords: 0211 other engineering and technologies; 02 engineering and technology

In academia and industries, graph neural networks (GNNs) have emerged as a powerful approach to data processing ranging from node classification link prediction tasks clustering tasks. GNN models are usually handcrafted. However, building handcrafted is difficult requires expert experience because model components complex sensitive variations. The complexity of has brought significant challenges the existing efficiencies GNNs. Hence, many studies focused on automated machine learning frameworks search for best targeted this work, we provide comprehensive review automatic summarize status field facilitate future progress. We categorize into three dimensions according them. After reviewing representative works each dimension, discuss promising research directions in rapidly growing field.