Light PhD Seminar: An Introduction to Algorithmic Fairness

Location: Maryam Mirzakhani Seminar Room at BCAM

Title
An Introduction to Algorithmic Fairness

Abstract
Automated decision-making tools are increasingly being used in high-stake applications involving human lives, ranging from healthcare, hiring, criminal incarceration, and so forth. These algorithms have been found to reproduce discriminating behaviors with respect to the gender, race, age, sexual orientation etc. of individuals. In the light of the severe impact that such potential harm has in the society, the research field of algorithmic fairness has grown substantially. Indeed, the scientific community has come up with a diverse pool of mathematical definitions to measure the fairness guarantees of machine learning models, as well as fairness-enhancing interventions that aim to improve the aforesaid guarantees. This LIGHT seminar aims to provide an overview regarding the key notions, definitions and approaches that form the basis of the growing research field denoted as algorithmic fairness.