DeepGG: Deep Graph Generator

Learning a state-based generative model of graph distributions.

@inproceedings{stier2020deep,
  title={DeepGG: a Deep Graph Generator},
  author={Stier, Julian and Granitzer, Michael},
  booktitle={Advances in Intelligent Data Analysis XIX: 19th International Symposium on Intelligent Data Analysis, IDA 2021, Porto, Portugal, April 26--28, 2021, Proceedings},
  pages={325},
  organization={Springer Nature}
}

Reproducing Experiments

• install the conda environment with conda env create -f environment.yml
• activate the environment conda activate sur-deepgg
• configure your (hyper)parameters (first ~20 variables)
• invoke as much as possible computations via python deepgg_pipeline.py
• merge the computations as shown in deepgg-merge-computations.ipynb
• have a look over the exemplary notebooks of how to visualize some aspects of the computed models