I didn’t find any example in the DGL document. But I want to use the graph auto-encoder algorithm to realize unsupervised learning. Maybe there are some examples in the document, can anybody give me a simple one to understand how is DGL working on graph auto-encoder?

Could you provide a reference paper?

Adversarially Regularized Graph Autoencoder for Graph Embedding

One2Multi Graph Autoencoder for Multi-view Graph Clustering

The No.1 paper demonstrates a classical and practical method.

Here is an implementation for Graph Autoencoder by Kipf & Welling in DGL. Maybe you will find it helpful.

Thanks for your reply, I have notice that repo, but I can not reimplement it due to the dataset structure.

Actually, I’m looking forward an official version of Graph Auto-encoder.

Can you elaborate more about the data structure issue?