/GGD

The source code of Generated Graph Detection.

Primary LanguagePythonMIT LicenseMIT

Generated Graph Detection

This is Pytorch implementation of Generated Graph Detection, as described in our paper:

Step 0: Datasets

The datasets used in our paper,

["AIDS", "alchemy_full", "deezer_ego_nets", "DBLP_v1", "github_stargazers", "COLLAB", "twitch_egos"]

All datasets can be obtained in https://chrsmrrs.github.io/datasets/

Step 1: Graph Generation

We follow the instruction on https://github.com/xiangsheng1325/GraphGenerator to generate graphs based on Graphite, GRAN, sbmgnn, vgae.

We follow the instruction on https://github.com/JiaxuanYou/graph-generation to generate graphs based on GraphRNN, GraphVAE.

The generated graphs based on ER and BA can be generated by the following commands:

python3 main.py -e Generate_ER -d AIDS
python3 main.py -e Generate_BA -d AIDS

Step 2: Sample dataset

python3 main.py -e sample_dataset -d AIDS

Step 3: Run experiments

python3 main.py -e train_models -d AIDS -o .pkl -s s1