Reproducibility Study Of Learning Fair Graph Representations Via Automated Data Augmentations
This is an incremental reproducibility study that addresses fairness in graph-based learning for researchers and practitioners, extending the method to new tasks.
This study reproduced and extended the Graphair framework for fair graph representations, partially validating one of three original claims and fully substantiating the other two, while showing comparable fairness-accuracy trade-offs for mixed dyadic-level fairness and superior trade-offs for subgroup dyadic-level fairness in link prediction tasks.
In this study, we undertake a reproducibility analysis of 'Learning Fair Graph Representations Via Automated Data Augmentations' by Ling et al. (2022). We assess the validity of the original claims focused on node classification tasks and explore the performance of the Graphair framework in link prediction tasks. Our investigation reveals that we can partially reproduce one of the original three claims and fully substantiate the other two. Additionally, we broaden the application of Graphair from node classification to link prediction across various datasets. Our findings indicate that, while Graphair demonstrates a comparable fairness-accuracy trade-off to baseline models for mixed dyadic-level fairness, it has a superior trade-off for subgroup dyadic-level fairness. These findings underscore Graphair's potential for wider adoption in graph-based learning. Our code base can be found on GitHub at https://github.com/juellsprott/graphair-reproducibility.