Graph generative models can be used to generate synthetic graphs for benchmarking and analysis tasks. 

For example, suppose you’ve developed a community detection algorithm. It would be reasonable to expect that your community detection approach should be able to infer the underlying communities in a graph generated by an SBM model.

University of Michigan - Ann Arbor

The primary insight in the common traditional models is that we specify a generation process or probability model, which allows us to capture some useful property of real-world graphs while still being tractable and easy to analyze. Historically, these traditional generation models have been used in two key applications:
- Generating synthetic data
- Creating null models

Traditional Applications

Hamilton, W.L. Graph Representation Learning. *Synthesis Lectures on Artificial Intelligence and Machine Learning*. 2020, 14(3), 1–159. https://doi.org/10.2200/S01045ED1V01Y202009AIM046

Graph Representation Learning by William Hamilton

Generating Synthetic Data

One key application task for traditional graph generation approaches is the creation of null models. In general, traditional generative models of graphs give us the ability to interrogate what sorts of graph characteristics can be easily explained by simple generative processes. In a statistical sense, they provide us with null models that we can use as reference points for our understanding of real-world graphs.

Learn Before

Related