Date: Monday, November 17th, 2025
10:00 am – 11:00 am Pacific Time
1:00 pm – 2:00 pm Eastern Time
Location: Weekly Seminar, Zoom
Title: Why we need new fairness metrics and how to use them
Abstract:
A rich set of formal definitions and statistical criteria has been developed in existing fairness research. Given that, some scholars argue that research should now focus more on providing normative ML pipelines to assist practitioners in making decisions, rather than proposing new computational methods for fairness. Despite the valuable starting points these criteria offer, they remain limited in important ways.
In this talk, I will present empirical illustrations of these limitations, introduce several new fairness measures, and discuss one way to use them to help construct fairer models. The talk will cover five of my recent works, with an emphasis on the overarching ideas. Technical details will be discussed selectively depending on time and audience interest.
Bio:
Yijun Bian is a Marie Skłodowska-Curie postdoctoral fellow at the Department of Computer Science, University of Copenhagen, where her work mainly focuses on fairness in machine learning. She received her Ph.D. degree from the University of Science and Technology of China, working on ensemble learning. Her research interests lie in reliable and trustworthy aspects of machine learning.