This repository includes the reproduction of Luo et al. 2024: Are Your Models Still Fair? Fairness Attacks on Graph Neural Networks via Node Injections.
Their original repo is found here
Our experiments can be replicated through all_runs.ipynb
Make sure to use a machine with a gpu and with python not obtained through any virtual environment
- Due to size limitation, the processed datasets are stored in google drive as
data.zip. The datasets include Pokec-z, Pokec-n and DBLP.
This is hosted by the original authors. We have included this zip file in our repository.
Make sure to copy the dataset bin files from the data folder togfair/dataset/andfairsin/data_nifa/