Reference Code for AAAI-20 paper "Multi-Stage Self-Supervised Learning for Graph Convolutional Networks on Graphs with Few Labels"

Please refer to https://arxiv.org/pdf/1902.11038.pdf to look into the details of our paper.

As for the official tensorflow code, please refer to Deeper Insights into Graph Convolutional Networks for Semi-Supervised Learning (AAAI 2018) github: https://github.com/liqimai/gcn/tree/AAAI-18/, since our code is directly adapted from this paper.

Due to my graduation issue and research interest transfer, I may not have enough time to recap and release the tensorflow code in a short period and I appologize for that. However, I found an unofficial pytorch implemtentation of our work: https://github.com/Junseok0207/M3S_Pytorch, which seems to be a nice implementation and shows similar results.

Please contact ajksunke@pku.edu.cn if you have any questions I can help. More information can be found in Ke Sun's personal website: https://sites.google.com/view/kesun.

Reference

@inproceedings{sun2020multi,
  title={Multi-Stage Self-Supervised Learning for Graph Convolutional Networks on Graphs with Few Labeled Nodes.},
  author={Sun, Ke and Lin, Zhouchen and Zhu, Zhanxing},
  booktitle={AAAI},
  pages={5892--5899},
  year={2020}
}