This session is designed to not only raise awareness of harmful biases in machine learning, but also to equip attendees with practical tools to measure and mitigate them effectively. We will begin by exploring the origins and impacts of bias– tracing its roots from societal influences to manifestations in AI systems. Attendees will gain a deep understanding of the many forms of bias, how it shapes model outcomes, and why it is essential to address it early and often in the ML lifecycle.

In the second half, we will transition from theory to practice. Participants will engage in a hands-on session, working with a real-world dataset to measure and identify bias within it. We will explore various pre-processing methods to understand how decisions in this modeling stage can introduce or amplify bias, and learn how to apply techniques that mitigate these effects. By the end of both sessions, participants will not only have a wider understanding of the field of AI Fairness, but also walk away with the tools and knowledge to build more responsible and trustworthy machine learning systems.