Mitigating Bias in Set Selection with Noisy Protected Attributes

The course aims to teach learners how to mitigate bias in set selection when dealing with noisy protected attributes. The learning outcomes include understanding subset selection, fair subset selection, handling missing protected attributes, noise modeling, denoising problems, simulating scenarios, evaluating fairness metrics, and drawing conclusions. The teaching method involves a research track presentation at the FAccT 2021 conference. The intended audience for this course is individuals interested in addressing bias in data selection processes, particularly in the context of protected attributes.