All Tools
F
DataFreeOpen Source
FAIRLEARN
Assess and mitigate ML fairness issues
MIT
ABOUT
Production models can treat groups differently even when accuracy looks fine. Fairlearn measures disparity with group fairness metrics, applies mitigation algorithms at train or post-process time, and produces reports so teams can audit models before they ship harmful decisions.
INSTALL
pip install fairlearnINTEGRATION GUIDE
1. Measure group fairness metrics on a classification or ranking model
2. Apply reduction or post-processing mitigations to reduce disparity
3. Generate fairness reports for model review and compliance workflows
TAGS
pythonfairnessresponsible-aimetricsmitigationmachine-learning