Fairer and more accurate, but for whom?

Abstract

Complex statistical machine learning models are increasingly being used or considered for use in high-stakes decision-making pipelines in domains such as financial services, health care, criminal justice and human services. These models are often investigated as possible improvements over more classical tools such as regression models or human judgement. While the modeling approach may be new, the practice of using some form of risk assessment to inform decisions is not. When determining whether a new model should be adopted, it is therefore essential to be able to compare the proposed model to the existing approach across a range of task-relevant accuracy and fairness metrics. Looking at overall performance metrics, however, may be misleading. Even when two models have comparable overall performance, they may nevertheless disagree in their classifications on a considerable fraction of cases. In this paper we introduce a model comparison framework for automatically identifying subgroups in which the differences between models are most pronounced. Our primary focus is on identifying subgroups where the models differ in terms of fairness-related quantities such as racial or gender disparities. We present experimental results from a recidivism prediction task and a hypothetical lending example.

arXiv commentPresented as a poster at the 2017 Workshop on Fairness, Accountability, and Transparency in Machine Learning (FAT/ML 2017)

References in corpus (2)

• Identifying Significant Predictive Bias in Classifiers
• Fairness in Criminal Justice Risk Assessments: The State of the Art

Cited by in corpus (16)

• Prediction-Based Decisions and Fairness: A Catalogue of Choices, Assumptions, and Definitions
• Fairness in Machine Learning: A Survey
• Distill-and-Compare: Auditing Black-Box Models Using Transparent Model Distillation
• Achieving Fairness through Adversarial Learning: an Application to Recidivism Prediction
• Calibration for the (Computationally-Identifiable) Masses
• A survey on datasets for fairness-aware machine learning
• What's in a Name? Reducing Bias in Bios without Access to Protected Attributes
• Investigating Human + Machine Complementarity for Recidivism Predictions
• Characterizing Fairness Over the Set of Good Models Under Selective Labels
• SoK: Machine Learning Governance
• To Split or Not to Split: The Impact of Disparate Treatment in Classification
• Locating disparities in machine learning
• Reasons, Values, Stakeholders: A Philosophical Framework for Explainable Artificial Intelligence
• Fairness Sample Complexity and the Case for Human Intervention
• Automated Dependence Plots
• Individual Fairness in Hindsight