Fairness Beyond Disparate Treatment & Disparate Impact: Learning Classification without Disparate Mistreatment

Abstract

Automated data-driven decision making systems are increasingly being used to assist, or even replace humans in many settings. These systems function by learning from historical decisions, often taken by humans. In order to maximize the utility of these systems (or, classifiers), their training involves minimizing the errors (or, misclassifications) over the given historical data. However, it is quite possible that the optimally trained classifier makes decisions for people belonging to different social groups with different misclassification rates (e.g., misclassification rates for females are higher than for males), thereby placing these groups at an unfair disadvantage. To account for and avoid such unfairness, in this paper, we introduce a new notion of unfairness, disparate mistreatment, which is defined in terms of misclassification rates. We then propose intuitive measures of disparate mistreatment for decision boundary-based classifiers, which can be easily incorporated into their formulation as convex-concave constraints. Experiments on synthetic as well as real world datasets show that our methodology is effective at avoiding disparate mistreatment, often at a small cost in terms of accuracy.

To appear in Proceedings of the 26th International World Wide Web Conference (WWW), 2017. Code available at: https://github.com/mbilalzafar/fair-classification

References in corpus (2)

• Equality of Opportunity in Supervised Learning
• Inherent Trade-Offs in the Fair Determination of Risk Scores

Cited by in corpus (23)

• Fairness in Machine Learning
• Explaining Models: An Empirical Study of How Explanations Impact Fairness Judgment
• User-oriented Fairness in Recommendation
• Fighting Fire with Fire: Using Antidote Data to Improve Polarization and Fairness of Recommender Systems
• Fairness in Algorithmic Decision Making: An Excursion Through the Lens of Causality
• Soliciting Stakeholders' Fairness Notions in Child Maltreatment Predictive Systems
• Fairness in Semi-supervised Learning: Unlabeled Data Help to Reduce Discrimination
• Case Study: Predictive Fairness to Reduce Misdemeanor Recidivism Through Social Service Interventions
• Fair Classification with Group-Dependent Label Noise
• Trajectories of Blocked Community Members: Redemption, Recidivism and Departure
• Towards Individual and Multistakeholder Fairness in Tourism Recommender Systems
• Visus: An Interactive System for Automatic Machine Learning Model Building and Curation
• Out of Context: Investigating the Bias and Fairness Concerns of "Artificial Intelligence as a Service"
• Runtime Monitoring of Dynamic Fairness Properties
• Towards Fairness in Personalized Ads Using Impression Variance Aware Reinforcement Learning
• A Framework for Designing Fair Ubiquitous Computing Systems
• Evaluating the Fairness of Discriminative Foundation Models in Computer Vision
• RecRec: Algorithmic Recourse for Recommender Systems
• Improving Fairness in Adaptive Social Exergames via Shapley Bandits
• Fair Inputs and Fair Outputs: The Incompatibility of Fairness in Privacy and Accuracy
• Convex Fairness Constrained Model Using Causal Effect Estimators
• Long-Term Fairness in Sequential Multi-Agent Selection with Positive Reinforcement
• On the Impact of Data Quality on Image Classification Fairness