In this presentation, we will do a recap of the core concepts of a neural network after which the essential functions that make up a graph neural network will be described. We will show how the message passing framework that is at the core of every graph neural network operates and describe the different variations that make up the most common graph neural network models. We will illustrate an example of such a graph neural network; the Graph Attention Network, and how this network can be integrated into a Neo4j workflow to perform node classification predictions. We will finish off with a brief description of the core aspects of Knowledge graphs and why Graph Neural Networks are important to Knowledge graph completion.
The "Continuous Learning" series is Neo4j staff presenting to Neo4j staff, an internal presentation to which community members are now invited to join.
- Add the event to your calendar
- Join us on Neo4j Discord #series to get connection details and to chat during the presentation