Reproduce “Characterizing Predictive Multiplicity in Ranking and Its Fairness Implications” 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
Ainhize Barrainkua, Mikel Malagón
The authors define the model-agnostic Rashomon set of top-K candidate groups within ε of the best set’s total score. They use it to measure inclusion instability and find sets with lower representation disparity. Across tested datasets, less disparate sets are sometimes found at ε = 0, with no loss in score-based utility. This concerns the paper’s representation-based fairness measure.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.