1.2 Applications of Graph ML
1.2.1 Classic Graph ML Tasks
•
Node Classification
•
Link Prediction
•
Graph Classification (ex. Molecule property prediction)
•
Clustering
•
Graph generation(drug discovery), evolution(physical simulation)
•
Node-level ML Tasks
â—¦
Protein Folding (DeepMind’s AlphaFold)
•
Edge-level ML Tasks
â—¦
Recommender Systems
â—¦
Drug Side Effects (when taking multiple drugs)
â—¦
Traffic Prediction
•
Graph-level ML Tasks
â—¦
Drug discovery
â—¦
Physics Simulation
1.2.2 Choice of Graph Representation
•
Directed vs Undirected Graphs
•
Node degrees
•
Bipartite graph
â—¦
two different types of nodes
â—¦
where nodes only interact with the other type of node
â—¦
Folded/Projected Bipartite Graphs
â–ª
Bipartite graph → Projection graph
•
Representing Graphs: Adjacency Matrix
â—¦
In real-world matrix, adjacency matrices are extremely sparse
•
Representing Graphs: Edge list
â—¦
Easier to work with large & sparse graph
â—¦
Allow us to quickly retrieve all neighbors of a given node
•
Node and Edge Attributes
•
More Types of Graphs
â—¦
Unweighted and Weighted
â—¦
Self-edges (self-loops)
â—¦
Multigraph
•
Connectivity
â—¦
Connected undirected graph
â—¦
Connected directed graph
â–ª
A strongly connected directed graph
â–ª
A weakly connected directed graph (disregard the edge directions)