Speaker
Description
// Level: BEGINNER ///
In this talk we survey the role of machine learning methods in questions of social justice and discrimination. First, we take a bird’s eye view on which domains may be particularly affected, how machine learning can sustain or even promote inequalities, and whether there are also opportunities for ML to help reduce or prevent discrimination in practice. Via a deep dive into automated data-driven decision-making in consequential scenarios, we learn about the interactions of technical aspects with societal questions and introduce a broader perspective of the life-cycle of ML methods. Throughout, we try to give concrete examples of ML models arguably acting “unfair” and try to distill potential mindsets and techniques to avoid such failure modes in the future.