Peizhao Li, Yifei Wang, Han Zhao, Pengyu Hong, and Hongfu Liu. "On Dyadic Fairness: Exploring and Mitigating Bias in Graph Connections , ICLR 2021.
Disparate impact has raised serious concerns in machine learning applications and its societal impacts. In response to the need of mitigating discrimination, fairness has been regarded as a crucial property in algorithmic designs. In this work, we study the problem of disparate impact on graph-structured data. Specifically, we focus on dyadic fairness, which articulates a fairness concept that a predictive relationship between two instances should be independent of the sensitive attributes. Based on this, we theoretically relate the graph connections to dyadic fairness on link predictive scores in learning graph neural networks, and reveal that regulating weights on existing edges in a graph contributes to dyadic fairness conditionally. Subsequently, we propose our algorithm, FairAdj, to empirically learn a fair adjacency matrix with proper graph structural constraints for fair link prediction, and in the meanwhile preserve predictive accuracy as much as possible. Empirical validation demonstrates that our method delivers effective dyadic fairness in terms of various statistics, and at the same time enjoys a favorable fairness-utility tradeoff.
@inproceedings{
li2021on,
title={On Dyadic Fairness: Exploring and Mitigating Bias in Graph Connections},
author={Peizhao Li and Yifei Wang and Han Zhao and Pengyu Hong and Hongfu Liu},
booktitle={International Conference on Learning Representations},
year={2021},
url={https://openreview.net/forum?id=xgGS6PmzNq6}
}
Please check or adjust the environment configurations in req.txt for PyTorch since the CUDA version may differ.
conda create -n fairadj --file req.txt
conda activate fairadj
Data for Cora and Citeseer are stored in ./data.
Execute python file in ./src:
cd src
Experiments on FairAdj:
python main.py --dataset "cora" --device "cuda:0" --eta 0.2 --T2 20
python main.py --dataset "citeseer" --device "cuda:0" --eta 5. --T2 20
You can set T2 to various values, see Figure 1 and 6 in our paper for more results.
Experiments on Adversarial Graph Embedding:
python adv.py --dataset "cora" --device "cuda:0" --lr 0.001 --alpha 10.
python adv.py --dataset "citeseer" --device "cuda:0" --lr 0.001 --alpha 10.
Similarly, alpha can be set to various values, see Figure 1 and 6 in our paper for more results.
We provide all the experimental results and figures in ./res.