Explore a conference talk on the cost of fairness in binary classification, presented by Robert C. Williamson at FAT* 2018. Delve into the award-winning research conducted by Williamson and Aditya Krishna Menon from The Australian National University and Data61. Learn about key concepts such as binary classification, loss functions, risk analysis, and fairness-aware settings. Examine the optimal classifier, fairness frontier, and interpretations of the findings. Gain insights into the partial truth of stereotypes and their implications in machine learning algorithms. Understand the technical contributions that earned this presentation the Best Paper award for Best Technical Contribution at the conference.
Overview
Syllabus
Introduction
Binary classification
Loss function
Why bother
Formalization
Risk
Analysis
Fairness aware setting
Optimal classifier
Fairness frontier
Interpretations implications
Partial truth of stereotypes
Taught by
ACM FAccT Conference
