Reproduce “Rethinking Fairness Metrics for Multi-Class and Multi-Group Assessment: Not Everything is Binary” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Matteo Grazioso, Fabio Naccarella, Chiara Gallese, Marco S. Nobile
The authors present FairMetrics, a framework for multi-class fairness audits across intersecting sensitive groups. It measures disparities without requiring one-versus-rest reductions. In the Students’ Dropout and Academic Success dataset, the same model’s Enrolled Conditional Demographic Parity score is 0.7090 natively and 0.4197 after prediction binarization, a −0.2894 gap; this empirical difference is not proof of improved fairness.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.