Track

Graph ML

Goal: Learn representation learning on graphs from node embeddings through message-passing GNNs, knowledge-graph models, graph generation, and graph transformers.

Prereqs: Deep Learning. ML Basics helps for the embedding and classification framing.

Status: done

Work through the steps in order. Bold links open YouTube.

Step	Concept	YouTube	Read
1	Why graphs for machine learning	Stanford CS224W — Why Graphs	Distill — A Gentle Introduction to Graph Neural Networks
2	Convolutions on graphs		Distill — Understanding Convolutions on Graphs
3	Random-walk node embeddings	Graph Embeddings (node2vec) explained	Grover & Leskovec 2016 — node2vec: Scalable Feature Learning for Networks
4	Message passing neural networks	Simple Message Passing on Graphs	Gilmer et al. 2017 — Neural Message Passing for Quantum Chemistry
5	Graph convolutional networks	Graph Convolutional Networks (GCN) \| GNN Paper Explained	Kipf & Welling 2017 — Semi-Supervised Classification with Graph Convolutional Networks
6	Inductive learning with GraphSAGE	GraphSAGE: Inductive Representation Learning on Large Graphs	Hamilton et al. 2017 — Inductive Representation Learning on Large Graphs
7	Graph attention networks	GAT: Graph Attention Networks	Veličković et al. 2018 — Graph Attention Networks
8	Expressive power and GIN	Weisfeiler-Lehman test and message-passing NNs	Xu et al. 2019 — How Powerful are Graph Neural Networks?
9	Knowledge graph embeddings	TransE — Translating Embedding for Knowledge Graphs	Bordes et al. 2013 — Translating Embeddings for Modeling Multi-relational Data
10	Deep generative models for graphs		You et al. 2018 — GraphRNN: Generating Realistic Graphs with Deep Auto-regressive Models
11	Graph transformers	Graphormer — Do Transformers Really Perform Bad for Graph Representation?	Ying et al. 2021 — Do Transformers Really Perform Bad for Graph Representation?

Found a broken link or an unclear step? Report a problem with this track.