chapter six

6 Graph Autoencoders

 

This chapter covers

  • Distinguishing between discriminative and generative models
  • Applying AutoEncoders (AEs) and Variational AutoEncoders (VAEs) to graphs
  • Building both types of Graph AutoEncoders (GAEs) with PyTorch Geometric
  • Performing link prediction with GAEs

So far, we have covered how classical deep learning architectures can be extended to work on graph-structured data. In Chapter 4 we considered convolutional GNNs, which apply the convolutional operator to identify patterns within the data. In Chapter 5, we explored the attention mechanism and how this can be used to improve performance for graph-learning tasks such as node classification.

6.1 Generative models: Learning how to generate

6.2 Deep generative model: Autoencoders

6.2.1 Applying a discriminative model to the MNIST dataset

6.2.2 Applying PCA to the MNIST dataset

6.2.3 Applying a generative model to the MNIST data

6.2.4 Building a Graph AutoEncoder for Link Prediction

6.3 Variational AutoEncoders

6.3.1 Applying a VAE to the MNIST data

6.3.2 Variational Graph AutoEncoders

6.4 Summary

6.5 References