When majority rules, minority loses: bias amplification of gradient descent
Journal:
arXiv
Published Date:
May 19, 2025
Abstract
Despite growing empirical evidence of bias amplification in machine learning,
its theoretical foundations remain poorly understood. We develop a formal
framework for majority-minority learning tasks, showing how standard training
can favor majority groups and produce stereotypical predictors that neglect
minority-specific features. Assuming population and variance imbalance, our
analysis reveals three key findings: (i) the close proximity between
``full-data'' and stereotypical predictors, (ii) the dominance of a region
where training the entire model tends to merely learn the majority traits, and
(iii) a lower bound on the additional training required. Our results are
illustrated through experiments in deep learning for tabular and image
classification tasks.
