Robust Fairness Under Covariate Shift
Abstract
Making predictions that are fair with regard to protected attributes (race, gender, age, etc.) has become an important requirement for classification algorithms. Existing techniques derive a fair model from sampled labeled data relying on the assumption that training and testing data are identically and independently drawn (iid) from the same distribution. In practice, distribution shift can and does occur between training and testing datasets as the characteristics of individuals interacting with the machine learning system change. We investigate fairness under covariate shift, a relaxation of the iid assumption in which the inputs or covariates change while the conditional label distribution remains the same. We seek fair decisions under these assumptions on target data with unknown labels. We propose an approach that obtains the predictor that is robust to the worst-case testing performance while satisfying target fairness requirements and matching statistical properties of the source data. We demonstrate the benefits of our approach on benchmark prediction tasks.
Additional Information
© 2021 Association for the Advancement of Artificial Intelligence. Published 2021-05-18. This work was supported by the National Science Foundation Program on Fairness in AI in collaboration with Amazon under award No. 1939743.Attached Files
Published - 17135-Article_Text-20629-1-2-20210518.pdf
Submitted - 2010.05166.pdf
Files
Name | Size | Download all |
---|---|---|
md5:a6025e981ce17c8aab07bd798ccdf0d1
|
2.9 MB | Preview Download |
md5:db91f7d12a5dfd997d75ac14761739d0
|
834.0 kB | Preview Download |
Additional details
- Eprint ID
- 111442
- Resolver ID
- CaltechAUTHORS:20211014-173153985
- IIS-1939743
- NSF
- Created
-
2021-10-14Created from EPrint's datestamp field
- Updated
-
2023-06-02Created from EPrint's last_modified field