Zoom (Talk)
Virtual Talk with Christoph Kern
Abstract
Prediction algorithms have become widespread in domains where poli- cies dictate resource allocation, access, and individual benefits. However, real-world applications of algorithmic decision-making (ADM) often reveal significant shortcomings, including ethical issues of algorithmic bias and fair- ness. The social impacts of ADM systems critically depend on the interaction between biases in training data and the decisions made during a systems’ de- sign, implementation, and evaluation. In this talk, we highlight how a critical social (data) science perspective can identify data practices that contribute to model unfairness and can offer methodology to advance reliability in algorithmic decision-making. We first demonstrate how biases in training data can be mitigated or reinforced along the modeling pipeline dependent on preprocessing, analysis and evaluation decisions. We further present a set of common, yet unreflective data prac- tices which disproportionately affect minority groups and can distort model comparisons. We next highlight how methodology from the social sciences in interaction with participatory AI approaches can democratize model design decisions and systematically assess their fairness implications. We further present ingredients of an expanded ADM toolkit which is critically needed for reliable data-driven decision-making in dynamic social environments.
Bio
Christoph Kern is Junior Professor of Social Data Science and Statistical Learning at the Ludwig-Maximilians-University of Munich, Project Director at the Mannheim Centre for European Social Research (MZES) and Research Assistant Professor at the Joint Program in Survey Methodology (JPSM) at the University of Maryland. His research focuses on the social impacts of algorithmic decision-making and on methodology to mitigate algorithmic unfairness and improve training data quality.
Zoom Registration
The Talk will take place on Wednesday, December 11th, 13:00 CEST.
Please register with office@c3s.uni-frankfurt.de to receive the Zoom login.