Mitigating Bias in Set Selection with Noisy Protected Attributes

Explore a 16-minute conference talk from the FAccT 2021 virtual event that delves into the challenge of mitigating bias in set selection when dealing with noisy protected attributes. Learn about subset selection, fair subset selection, and the complexities of missing protected attributes. Understand the noise model and denoising problem presented by the researchers. Examine the simulation and fairness metric used in the study. Gain insights into the importance of addressing bias in algorithmic decision-making processes, particularly when working with incomplete or imperfect data.

Introduction
Subset Selection
Fair Subset Selection
Missing Protected Attributes
Noise Model
Denoise Problem
Simulation
Fairness Metric
Conclusion