Variational-Graph-Autoencoder-learning-the-decoding-part

  • University of Trento
  • Advanced Machine Learning and Optimization Course by Prof.Passerini
  • Ali Akay [224414]
  • MSc. Artificial Intelligence Systems

Objective

The project aim to invastigate the application for the decoder for VGAE.It can be seen that inner product still have good result when we compare the models for scalability and complexity.On the other hand,there has been better applications in order to improve the performance of the decoder using flexiable generative models and sequence models. I did not see a significant result in outcomes by using MLP in the decoder. However, I would like to try this experiment again on VGAE with sequence models for decoder in order to get better results.