Log In Sign Up

A Maximal Correlation Approach to Imposing Fairness in Machine Learning

by   Joshua Lee, et al.

As machine learning algorithms grow in popularity and diversify to many industries, ethical and legal concerns regarding their fairness have become increasingly relevant. We explore the problem of algorithmic fairness, taking an information-theoretic view. The maximal correlation framework is introduced for expressing fairness constraints and shown to be capable of being used to derive regularizers that enforce independence and separation-based fairness criteria, which admit optimization algorithms for both discrete and continuous variables which are more computationally efficient than existing algorithms. We show that these algorithms provide smooth performance-fairness tradeoff curves and perform competitively with state-of-the-art methods on both discrete datasets (COMPAS, Adult) and continuous datasets (Communities and Crimes).


page 1

page 2

page 3

page 4


Fast Fair Regression via Efficient Approximations of Mutual Information

Most work in algorithmic fairness to date has focused on discrete outcom...

Gradual (In)Compatibility of Fairness Criteria

Impossibility results show that important fairness measures (independenc...

A continuous framework for fairness

Increasingly, discrimination by algorithms is perceived as a societal an...

Optimal algorithms for group distributionally robust optimization and beyond

Distributionally robust optimization (DRO) can improve the robustness an...

Fairness-Aware Neural Réyni Minimization for Continuous Features

The past few years have seen a dramatic rise of academic and societal in...

What About Applied Fairness?

Machine learning practitioners are often ambivalent about the ethical as...

Calibrated Data-Dependent Constraints with Exact Satisfaction Guarantees

We consider the task of training machine learning models with data-depen...