FawNA

This repository contains implementations of the software developed for the Fairness Aware Algorithms for Network Analysis(FAI) project, which is jointly funded by the National Science Foundation and Amazon under the Fairness in AI program. The goal of this project is to develop fairness-aware algorithms that can maintain the high utility of decisions generated by the AI systems without discriminating against particular subgroups of the population. Specifically, this research will address fundamental issues of fairness in algorithms that utilize network data in their decision making (More Information).

Project Highlights:

• Fairness-Aware Network Link Prediction[1]: This work examines the filter bubble problem from the perspective of algorithm fairness by introducing (1) a dyadic-level fairness criterion based on network modularity measure and (2) a novel framework that combines adversarial network representation learning with supervised link prediction to alleviate the filter bubble problem.

• Fairness Perception[2]: This work investigates the issue of algorithmic fairness from a network-centric perspective by introducing a notion of fairness perception.

• Fairness-Aware Node Classification[3]: In this work, we identified two potential biases that may hamper the effectiveness of existing graph neural network (GNN) formulation, namely, (1) bias towards the majority class when classifying nodes with an imbalanced class distribution and (2) bias against the protected group based on some predefined sensitive attributes. The proposed framework called Fairness-Aware Cost Sensitive Graph Convolutional Network (FACS-GCN) is designed to overcome both types of biases.

• Fairness-Aware Graph Sampling[4]: This work focuses on the development of a graph sampling algorithm that considers the tradeoff between preserving fairness and the topological properties of a network.

Publications:

1. Farzan Masrour, Tyler Wilson, Heng Yan, Pang-Ning Tan, Abdol-Hossein Esfahanian. Bursting the Filter Bubble: Fairness-Aware Network Link Prediction. In Proceedings of the 34th AAAI Conference on Artificial Intelligence (AAAI-2020), New York, NY (2020).

2. Farzan Masrour, Pang-Ning Tan, and Abdol-Hossein Esfahanian. Fairness Perception from a Network-Centric Perspective. In Proceedings of the 20th IEEE International Conference on Data Mining, Sorrento, Italy (2020).

3. Francisco Santos, Junke Ye, Farzan Masrour, Pang-Ning Tan, and Abdol-Hossein Esfahanian. FACS-GCN: Fairness-Aware Cost-Sensitive Boosting of Graph Convolutional Networks In Proc of IEEE International Joint Conference on Neural Networks (IJCNN-2022), Padova, Italy (2022).

4. Farzan Masrour, Francisco Santos, Pang-Ning Tan, and Abdol-Hossein Esfahanian. Fairness-Aware Graph Sampling for Network Analysis In Proc of the 22nd IEEE International Conference on Data Mining, Orlando, FL (2022).