Equalizing Financial Impact in Supervised Learning
This work addresses fairness in financial decisions for marginalized groups, but it is incremental as it builds on an existing criterion.
The paper tackles the limitation of existing fairness criteria in supervised learning by proposing 'equalized financial impact' to address disproportionate financial harm across protected groups in loan decisions, modifying a prior fairness criterion.
Notions of "fair classification" that have arisen in computer science generally revolve around equalizing certain statistics across protected groups. This approach has been criticized as ignoring societal issues, including how errors can hurt certain groups disproportionately. We pose a modification of one of the fairness criteria from Hardt, Price, and Srebro [NIPS, 2016] that makes a small step towards addressing this issue in the case of financial decisions like giving loans. We call this new notion "equalized financial impact."