pregel

Tensorflow implementation of Graph Convolutional Network

Setup

• sudo pip3 install -r requirements.txt

Run

• python3 main.py -h to view all the config parameters. Update the default parameters in the main.py file.
• python3 main.py to run the models.
• tensorboard --logdir=run1:<tensorboard-dir> -port 6006 to run tensorboard and go to http://localhost/6006

References

This work is an attempt to reproduce some of the works related to graph convolutional networks:

Semi-Supervised Classification with Graph Convolutional Networks

@article{kipf2016semi,
  title={Semi-Supervised Classification with Graph Convolutional Networks},
  author={Kipf, Thomas N and Welling, Max},
  journal={arXiv preprint arXiv:1609.02907},
  year={2016}
}

• Author's implementation in Tensorflow
• Author's implementation in Keras
• Author's implementation in PyTorch

Variational Graph Auto-Encoders

@article{hamilton2017representation,
  title={Representation Learning on Graphs: Methods and Applications},
  author={Hamilton, William L and Ying, Rex and Leskovec, Jure},
  journal={arXiv preprint arXiv:1709.05584},
  year={2017}
}

• Author's implementation in Tensorflow