Facets of Responsible Machine Learning
Flavio du Pin Calmon Assistant Professor, Harvard
This talk overviews recent results in two aspects of fair machine learning. First, we introduce a post-processing technique, "FairProjection," designed to ensure group fairness in prediction and classification. This method applies to any classifier without requiring retraining and attains state-of-the-art performance in both accuracy and group fairness assurance in probabilistic classification. We also present converse results based on Blackwell's "comparison of experiments" theorem that capture the limits of group-fairness assurance in classification. Second, we overview the concept of predictive multiplicity in machine learning. Predictive multiplicity arises when different classifiers achieve similar average performance for a specific learning task, yet produce conflicting predictions for individual samples.
