Simple and near-optimal algorithms for hidden stratification and multi-group learning Conference Paper

  


Authors: Tosh, C.; Hsu, D.
Title: Simple and near-optimal algorithms for hidden stratification and multi-group learning
Conference Title: 39th International Conference on Machine Learning (ICML)
Abstract: Multi-group agnostic learning is a formal learning criterion that is concerned with the conditional risks of predictors within subgroups of a population. The criterion addresses recent practical concerns such as subgroup fairness and hidden stratification. This paper studies the structure of solutions to the multi-group learning problem, and provides simple and near-optimal algorithms for the learning problem.
Journal Title Proceedings of Machine Learning Research
Volume: 162
Conference Dates: 2022 Jul 17-23
Conference Location: Baltimore, MD
ISBN: 2640-3498
Publisher: Journal Machine Learning Research  
Date Published: 2022-01-01
Start Page: 21633
End Page: 21657
Language: English
ACCESSION: WOS:000900130202031
PROVIDER: wos
Notes: Proceedings Paper -- Source: Wos