Document worth reading: “Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions”
A present flurry of evaluation train has tried to quantitatively define ‘fairness’ for choices based mostly totally on statistical and machine finding out (ML) predictions. The quick progress of this new topic has led to wildly inconsistent terminology and notation, presenting a essential downside for cataloguing and evaluating definitions. This paper makes an try to hold much-needed order. First, we explicate the numerous choices and assumptions made—usually implicitly—to justify the use of prediction-based choices. Next, we current how such choices and assumptions can improve issues about fairness and we present a notationally fixed catalogue of fairness definitions from the ML literature. In doing so, we offer a concise reference for contemplating by the options, assumptions, and fairness considerations of prediction-based selection strategies. Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions