Testing Group Fairness via Optimal Transport Projections
Nian Si1 Karthyek Murthy 2 Jose Blanchet1 Viet Anh Nguyen1 3
Abstract A natural first explanation for the reported algorithmic bi-
We present a statistical testing framework to de- ases is that the data used to train the algorithms may al-
tect if a given machine learning classifier fails to ready be corrupted by human biases (Buolamwini & Ge-
satisfy a wide range of group fairnes ...


雷达卡




京公网安备 11010802022788号







