Graph Autoencoders: A Journey from Start to End
Mohamed Abdel‐Basset, Nour Moustafa, Hossam Hawash, Zahir Tari · 2023
Graph autoencoders represent a family of graph intelligence methods of unsupervised learning on graph-structured data. They make use of latent variables and are able to learn interpretable latent representations for undirected graphs. This chapter deeply discusses and explains graph autoencoders that concentrate on customizing autoencoders to learn from graphs. The chapter begins by discussing the working mechanisms of standard autoencoder and how it is extended to graph data. Next, the second part of the chapter discusses and explains the viewpoint of variational graph autoencoders.