Description: Machine Learning (ML) is increasingly used to guide high-stakes decisions in various contexts such as college admissions, granting loans or hiring employees. By eliminating human judgment, ML-based decision-making promises to be neutral and objective and to find the right decisions in shorter time. At the same time, however, concerns are raised that algorithmic decision-making may foster discrimination and amplify existing biases that are fed into the models. This course discusses recent advances in the field of Fair ML: How can we measure algorithmic fairness? How can we mitigate biases and make ML models fair(er) and interpretable? A key aspect of the course will be linking fairness definitions and correction techniques with real world examples of automated decision-making.
Key: FairADM21
- Trainer/in: Christoph Kern
- Trainer/in: Frauke Kreuter