Equality of Opportunity in Supervised Learning
The authors of this paper use a case study of FICO credit scores to illustrate their notion that classification accuracy in supervised learning depends only on the joint statistics of the predictor and the protected attribute but not the interpretation of the individual features of the data. The study looks at the inherent limits of defining and identifying biases based on this notion and propose a criterion for producing discrimination against a specific sensitive attribute. They argue that one can optimally adjust any learned predictor to remove discrimination, according to their definition.
Abstract: We propose a criterion for discrimination against a specified sensitive attribute in supervised learning, where the goal is to predict some target based on available features. Assuming data about the predictor, target, and membership in the protected group are available, we show how to optimally adjust any learned predictor so as to remove discrimination according to our definition. Our framework also improves incentives by shifting the cost of poor classification from disadvantaged groups to the decision maker, who can respond by improving the classification accuracy. In line with other studies, our notion is oblivious: it depends only on the joint statistics of the predictor, the target and the protected attribute, but not on interpretation of individual features. We study the inherent limits of defining and identifying biases based on such oblivious measures, outlining what can and cannot be inferred from different oblivious tests. We illustrate our notion using a case study of FICO credit scores.
"Equality of Opportunity in Supervised Learning" by M. Hardt, E. Price, N. Sobrero Cornell University Library, [v1] 7 Oct 2016