Explore a virtual workshop panel discussion featuring computer scientists Michael Kearns and Cristopher Moore as they delve into the complex relationship between algorithms and social justice. Examine concepts such as group versus individual fairness, intersectionality, risk assessment, and pretrial detention. Gain insights into fairness notions in machine learning pipelines, fairness gerrymandering, and the trade-offs between fairness and accuracy. Analyze the barriers of meaning in algorithmic decision-making, focusing on the Arnold Public Safety Assessment and its implications. Investigate predictive policing, rearrest rates, and the evolving nature of algorithms in the justice system. Engage with thought-provoking questions addressed to both speakers, enhancing your understanding of the potential for algorithms to influence societal equity.
Overview
Syllabus
Introduction
Group fairness notions
Machine learning pipeline
Fairness gerrymandering
Intersectionality
Fairness vs Accuracy
Remarks
Discussion
Transparency
Barriers of Meaning
Arnold Public Safety Assessment
Barrier of Meaning
Study Results
Accuracy
Predictive Policing
What happens to the algorithm
Rearrest rates
Questions
Question for Chris
Question for Michael
Taught by
Santa Fe Institute
