I finished writing my intelligent textbook on Graph Neural Networks (GNNs).
Ship
This is a comprehensive, interactive intelligent textbook on Graph Neural Networks designed for graduate and advanced undergraduate students in machine learning, data mining, and network science. Built with MkDocs Material, it covers the full arc of the field — from graph theory and classical graph algorithms through modern architectures including graph transformers, knowledge graph models, and LLM+GNN integration.
Every chapter leads with a real-world motivating example, builds intuition before equations, provides full mathematical derivations, and includes complete runnable PyTorch Geometric code. 37 interactive MicroSims — p5.js simulations embedded directly in the browser — let you manipulate GNN internals hands-on: watch message passing animate across layers, tune attention weights in real time, and step through WL color refinement iteration by iteration.
The textbook is structured around a dependency-aware learning graph of 300 concepts, ensuring prerequisite sequencing is respected across all 27 chapters. Quizzes, glossary terms, and FAQs cover every major concept, with exercises distributed across all six levels of Bloom’s Taxonomy.
I am starting to make a Learn section
