Graph learning focuses on machine learning methods that operate on relational structure.
These methods are useful when data points are not independent, but are linked through networks, interactions, or spatial dependencies.
Typical examples include node classification, link prediction, graph embeddings, message passing, and graph neural networks.