Debiasing classifiers: is reality at variance with expectation?

2020/11/04 by Agrawal, Ashrya, Pfisterer, Florian, Bischl, Bernd +5 · 1 citation
#68Q32 #68T01 #68T05 #Computers and Society (cs.CY) #Econometrics (econ.EM) #FOS: Computer and information sciences #FOS: Economics and business #G.4 #I.2.0 #J.4 #Machine Learning (cs.LG)

paper · doi:10.48550/arxiv.2011.02407

Abstract

We present an empirical study of debiasing methods for classifiers, showing that debiasers often fail in practice to generalize out-of-sample, and can in fact make fairness worse rather than better. A rigorous evaluation of the debiasing treatment effect requires extensive cross-validation beyond what is usually done. We demonstrate that this phenomenon can be explained as a consequence of bias-variance trade-off, with an increase in variance necessitated by imposing a fairness constraint. Follow-up experiments validate the theoretical prediction that the estimation variance depends strongly on the base rates of the protected class. Considering fairness--performance trade-offs justifies the counterintuitive notion that partial debiasing can actually yield better results in practice on out-of-sample data.

Cited by

• Bias Begins with Data: The FairGround Corpus for Robust and Reproducible Research on Algorithmic Fairness

Related

• Towards Real-world Debiasing: Rethinking Evaluation, Challenge, and Solution
• Systematic Evaluation of Predictive Fairness
• When mitigating bias is unfair: multiplicity and arbitrariness in algorithmic group fairness
• Post-processing fairness with minimal changes
• Do Not Harm Protected Groups in Debiasing Language Representation Models