This is a short review on Graph Neural Networks done during a course given by Michal Valko for the Master's degree Mathematics, Vision, Learning (MVA) from ENS Paris-Saclay.
The project was advised by Claire Vernade and Rémy Degenne. Unlike other reviews, it tries to quickly present the concepts and recent advances around Graph Neural Networks, and focus on an example, the shortest path problem.