Spurred by recent advances, machine learning methods are beginning to prescribe decisions in high-stakes domains, including hiring and medical diagnoses. The standard machine learning paradigm that optimizes average-case performance, however, often yields models that perform poorly on tail subpopulations, such as underrepresented demographic groups. We present methods that optimize the worst-case subpopulation performance, thereby guaranteeing a uniform level of performance over all subpopulations. Our procedures are convex and computationally efficient. We prove finite-sample minimax upper and lower bounds, showing that uniform subpopulation performance comes at a cost in convergence rates. Empirically, our procedure improves performance on tail subpopulations, and provides a uniform level of performance through time by controlling latent minority proportions.
Bio:
Hongseok Namkoong is a Ph.D. candidate in the Department of Management Science & Engineering at Stanford University, where he is jointly advised by John Duchi and Peter Glynn. His research focuses on developing robust and reliable machine learning methods that can maintain a uniform level of performance under operation. He received his M.S. in statistics from Stanford, and B.S. in industrial engineering and mathematics from KAIST. Hong is a recipient of a number of awards and fellowships, including best paper awards at the Neural Information Processing Systems conference and the International Conference on Machine Learning (runner-up), and the best student paper award from the INFORMS Applied Probability Society.
