Reproducibility Study Of Learning Fair Graph Representations Via Automated Data Augmentations
Abstract: 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://anonymous.4open.science/r/graphair-reproducibility-2871.
Submission Length: Regular submission (no more than 12 pages of main content)
Changes Since Last Submission: We have improved results for the FairAdj baseline on the PubMed dataset. As communicated to reviewers, we planned to upload a new version including these results. The new results do not require any revisions to our claims. However, in this version, we also added an additional fairness-accuracy trade-off plot, now including both the $\Delta \text{DP}_m$ and $\Delta \text{DP}_s$ metrics. This allowed us to make a few revisions in Section 4.2, which we believe enhance the quality of our paper.
Assigned Action Editor: ~Guillaume_Rabusseau1
Submission Number: 2234
Loading