Leave-One-Out Unfairness

Explore the concept of leave-one-out unfairness in this 19-minute conference talk presented at FAccT 2021. Delve into the research findings of E. Black and M. Fredrikson as they examine the implications of this fairness metric in machine learning and algorithmic decision-making systems. Gain insights into how removing a single individual from a dataset can impact model outcomes and overall fairness. Understand the potential consequences of leave-one-out unfairness in real-world applications and learn about proposed methods to address this issue in algorithmic design and implementation.