A brief overview of sources of knowledge about graph neural networks
For a quick start in graph neural networks, I recommend that you familiarize yourself with the Stanford University course CS224W. Teacher Yuri Leskovets explains in detail and simply the processes of graph processing in machine learning on graphs, and also considers training graph neural networks. In addition to the video on YouTube, there are links to course materials that are located on the university website. Stanford CS224W: Machine Learning with Graphs and link to course materials on the university website.
There is also a book in this course called “ Graph Representation Learning” The book complements the video course and gives a deeper understanding of how GML and GNN work.
In addition, there is another book called Graph Neural Networks. I also recommend you read it, because it outlines popular methods of working with graph neural networks.
I also recommend that you familiarize yourself with the GitHub repository of this book. The repository contains all the source codes of programs that will help you quickly understand how it all works. link to the github repository from the book.
Here are a few Python libraries that will be useful for working with graphs: PyTorch Geometric, NetworkX, JSON and XML. Nebula Graph, Neo4j and MongoDB tools will be useful for storing graphs.