InFoRM: Individual Fairness on Graph Mining

This is a Python implementation of InFoRM: Individual Fairness on Graph Mining for the task of PageRank, spectral clustering and LINE, as described in our paper:

Jian Kang, Jingrui He, Ross Maciejewski, Hanghang Tong. InFoRM: Individual Fairness on Graph Mining. In Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 379-389. 2020 (KDD 2020).

Requirements

• python 3 (>3.7)
• numpy
• scipy
• sklearn
• networkx

Data

We provide data used in the paper in data folder. Have a look at the load_graph.py for your reference.

In the demos, we load PPI dataset.

Models

We provide three mutually exclusive debiasing method in method folder:

• debias_graph.py: Debiasing the input graph. Feel free to override __init__() and fit() functions to debias your own method.
• debias_model.py: Debiasing the mining model. Feel free to override __init__() and fit() functions to debias your own method.
• debias_result.py: Debiasing the mining results.

Demos

Please check our demos in demo_{#1}.ipynb where {#1} can be PageRank, spectral_clustering or LINE.

Reference

Please cite our paper if you use this code in your own work:

@inproceedings{kang2020inform,
  title={InFoRM: Individual Fairness on Graph Mining},
  author={Kang, Jian and He, Jingrui and Maciejewski, Ross and Tong, Hanghang},
  booktitle={Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining},
  pages={379–389},
  year={2020},
  organization={ACM}
}