Reductions for Fair Machine Learning

A python package that implements the black-box approach to fair classification described in the paper A Reductions Approach to Fair Classification.

Installation

The package can be installed via pip install fairlearn. To verify that it works, download test_fairlearn.py from the repository and run python test_fairlearn.py.

Instead of installing the package, you can clone the repository locally via git clone git@github.com:Microsoft/fairlearn.git. To verify that the package works, run python test_fairlearn.py in the root of the repository.

Usage

The function expgrad in the module fairlearn.classred implements the reduction of fair classification to weighted binary classification. Any learner that supports weighted binary classification can be provided as input for this reduction. Two common fairness definitions are provided in the module fairlearn.moments: demographic parity (class DP) and equalized odds (class EO). See the file test_fairlearn.py for example usage of expgrad.

Contributing

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.microsoft.com.

When you submit a pull request, a CLA-bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., label, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repositories using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.

Maintainers

fairlearn is maintained by:

• @MiroDudik

Releasing

If you are the current maintainer of this project:

1. Create a branch for the release: git checkout -b release-vxx.xx
2. Ensure that all tests return "ok": python test_fairlearn.py
3. Bump the module version in fairlearn/__init__.py
4. Make a pull request to Microsoft/fairlearn
5. Merge Microsoft/fairlearn pull request
6. Tag and push: git tag vxx.xx; git push --tags

Reporting Security Issues

Security issues and bugs should be reported privately, via email, to the Microsoft Security Response Center (MSRC) at secure@microsoft.com. You should receive a response within 24 hours. If for some reason you do not, please follow up via email to ensure we received your original message. Further information, including the MSRC PGP key, can be found in the Security TechCenter.