Fairlearn
A community-driven open-source Python toolkit to assess and mitigate fairness issues in AI systems.
Fairlearn is an open-source, community-driven project that helps data scientists improve the fairness of AI systems. Its Python package has fairness metrics and mitigation algorithms. The site adds guides and use cases, for example an assessment of a credit-card default model where groups defined by sex are compared and the disparities are mitigated. The project treats fairness as sociotechnical: the code is one part, and the context of use is the other. Install it from PyPI with pip install fairlearn.
- Price
- Free, open source
- Model
- free
- Platforms
- Python
- Category
- Bias and fairness testing
- Link checked
Published by AIGovernanceTools. How we check each AI governance tool.