Reproduce “Fairness Behind a Veil of Ignorance: A Welfare Analysis for Automated Decision Making” 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
Hoda Heidari, Claudio Ferrari, Krishna P. Gummadi, Andreas Krause
The authors propose ranking decision models by the expected utility of a subject behind a veil of ignorance. They use risk-averse utility of benefits and train models by minimizing prediction loss under a welfare threshold. In experiments, higher thresholds reduced Atkinson inequality, but fairness violations later rose in regression, so this is not a general guarantee of lower inequality.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.