Fairness Violations and Mitigation under Covariate Shift
arXiv:1911.00677 · doi:10.1145/3442188.3445865
Abstract
We study the problem of learning fair prediction models for unseen test sets distributed differently from the train set. Stability against changes in data distribution is an important mandate for responsible deployment of models. The domain adaptation literature addresses this concern, albeit with the notion of stability limited to that of prediction accuracy. We identify sufficient conditions under which stable models, both in terms of prediction accuracy and fairness, can be learned. Using the causal graph describing the data and the anticipated shifts, we specify an approach based on feature selection that exploits conditional independencies in the data to estimate accuracy and fairness metrics for the test set. We show that for specific fairness definitions, the resulting model satisfies a form of worst-case optimality. In context of a healthcare task, we illustrate the advantages of the approach in making more equitable decisions.
arXiv comment11 pages main and 7 pages supplementary, To appear at ACM FAccT '21, Previous arXiv version arXiv:1911.00677v1 was presented at Workshop on Fair ML for Health '19
References in corpus (2)
Cited by in corpus (7)
- Fairness Score and Process Standardization: Framework for Fairness Certification in Artificial Intelligence Systems
- What-is and How-to for Fairness in Machine Learning: A Survey, Reflection, and Perspective
- The Case for Globalizing Fairness: A Mixed Methods Study on Colonialism, AI, and Health in Africa
- Globalizing Fairness Attributes in Machine Learning: A Case Study on Health in Africa
- Cross-Institutional Transfer Learning for Educational Models: Implications for Model Performance, Fairness, and Equity
- Toward Operationalizing Pipeline-aware ML Fairness: A Research Agenda for Developing Practical Guidelines and Tools
- Algorithmic Fairness Datasets: the Story so Far