2.6 KiB
| layout | title |
|---|---|
| post | Contents |
These notes form a concise introductory course on machine learning with large-scale graphs. They mirror the topics topics covered by Stanford CS224W, and are written by the CS 224W TAs. {% include marginnote.html id='mn-construction' note='The notes are still under construction! They will be written up as lectures continue to progress. If you find any typos, please let us know, or submit a pull request with your fixes to our GitHub repository.'%}
You too may help make these notes better by submitting your improvements to us via GitHub. Note that submitting substantial improvements will result in bonus points being added to your overall grade!
Starting with the Fall 2019 offering of CS 224W, the course covers three broad topic areas for understanding and effectively learning representations from large-scale networks: preliminaries, network methods, and machine learning with networks. Subtopics within each area correspond to individual lecture topics.
Preliminaries
- Introduction and Graph Structure: Basic background for graph structure and representation
- Measuring Networks and Random Graphs: Network properties, random graphs, and small-world networks
- Motifs and Graphlets: Motifs, graphlets, orbits, ESU
Network Methods
- Structural Roles in Networks: RolX, Granovetter, the Louvain algorithm
- Spectral Clustering: Graph partitions and cuts, the Laplacian, and motif clustering
- Influence Maximization: Influential sets, submodularity, hill climbing
- Outbreak Detection: CELF, lazy hill climbing
- Link Analysis: PageRank and SimRank
- Network Effects and Cascading Behavior: Decision-based diffusion, probabilistic contagion, SEIZ
- Network Robustness: Power laws, preferential attachment
- Network Evolution: Densification, forest fire, temporal networks with PageRank
- Knowledge Graphs and Metapaths: Metapaths, reasoning and completion of KGs
Machine Learning with Networks
- Message Passing and Node Classification: Label propagation and collective classification
- Node Representation Learning: Shallow, DeepWalk, TransE, node2vec, and t-SNE
- Graph Neural Networks: GCN, SAGE, GAT
- Generative Models for Graphs: Variational Autoencoders, GraphRNN, Molecule GAN