Impartial Predictive Modeling: Ensuring Group Fairness in Arbitrary Models
Year: 2016
Type: preprint
Abstract: Fairness aware data mining (FADM) aims to prevent algorithms from discriminating against protected groups. The literature has come to an impasse as to what constitutes explainable variability as opposed to discrimination. We demonstrate that fairness is achieved by ensuring \emph{impartiality} with respect to sensitive characteristics and provide a framework for impartiality by accounting for different perspectives on the data generating process. In particular, fairness can only be precisely defined in a full-data scenario in which all covariates are observed. We then analyze how these models may be conservatively estimated via regression in partial-data settings. Decomposing the regression estimates provides insights into previously unexplored distinctions between explainable variability and discrimination. This framework yields a set of impartial estimates that are applicable in a wide variety of situations and post-processing tools to correct estimates from arbitrary models. This effectively separates prediction and fairness goals.
Source: arXiv (Cornell University)
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2496083297