Graph ML — Overview

Advanced Architectures — GNNs introduced message passing and the GCN/GraphSAGE/GAT architecture family. This section is the rest of the graph-ML picture: the task taxonomy those architectures actually solve, graph embedding methods that predate (and still compete with) GNNs, knowledge graphs, and how graph structure shows up in modern LLM systems via Graph RAG.

What's in this section

• Graph ML Fundamentals — graph representation, the task taxonomy (node classification, link prediction, graph classification), GIN, graph embeddings (node2vec, DeepWalk), knowledge graphs, and Graph RAG.

See the roadmap for the full path.