Fair Active Learning

TL;DR

This paper introduces fair active learning algorithms that balance model accuracy and demographic parity fairness, effectively reducing bias while minimizing labeling costs in societal applications.

Contribution

It presents novel algorithms for fair active learning that explicitly incorporate fairness constraints, particularly demographic parity, into the data selection process.

Findings

Effective reduction of bias in models through fair active learning

Maintains high accuracy while improving fairness metrics

Demonstrates success on benchmark datasets

Abstract

Machine learning (ML) is increasingly being used in high-stakes applications impacting society. Therefore, it is of critical importance that ML models do not propagate discrimination. Collecting accurate labeled data in societal applications is challenging and costly. Active learning is a promising approach to build an accurate classifier by interactively querying an oracle within a labeling budget. We design algorithms for fair active learning that carefully selects data points to be labeled so as to balance model accuracy and fairness. Specifically, we focus on demographic parity - a widely used measure of fairness. Extensive experiments over benchmark datasets demonstrate the effectiveness of our proposed approach.