Statistical Equity: A Fairness Classification Objective

TL;DR

This paper introduces a new fairness definition based on equity, aiming to address societal biases in machine learning by formalizing and operationalizing it for equitable classification, validated through various evaluations.

Contribution

It proposes a novel fairness concept rooted in equity, formalizes it, and demonstrates its practical utility in classification tasks with empirical validation.

Findings

Effective in reducing societal biases

Improves fairness in feedback loops

Validated through human and automatic evaluations

Abstract

Machine learning systems have been shown to propagate the societal errors of the past. In light of this, a wealth of research focuses on designing solutions that are "fair." Even with this abundance of work, there is no singular definition of fairness, mainly because fairness is subjective and context dependent. We propose a new fairness definition, motivated by the principle of equity, that considers existing biases in the data and attempts to make equitable decisions that account for these previous historical biases. We formalize our definition of fairness, and motivate it with its appropriate contexts. Next, we operationalize it for equitable classification. We perform multiple automatic and human evaluations to show the effectiveness of our definition and demonstrate its utility for aspects of fairness, such as the feedback loop.