Responsibly

.. image:: https://img.shields.io/badge/docs-passing-brightgreen.svg :target: https://docs.responsibly.ai

.. image:: https://img.shields.io/gitter/room/nwjs/nw.js.svg :alt: Join the chat at https://gitter.im/ResponsiblyAI/responsibly :target: https://gitter.im/ResponsiblyAI/responsibly

.. image:: https://img.shields.io/github/workflow/status/ResponsiblyAI/responsibly/CI/master.svg :target: https://github.com/ResponsiblyAI/responsibly/actions/workflows/ci.yml

.. image:: https://img.shields.io/coveralls/ResponsiblyAI/responsibly/master.svg :target: https://coveralls.io/r/ResponsiblyAI/responsibly

.. image:: https://img.shields.io/scrutinizer/g/ResponsiblyAI/responsibly.svg :target: https://scrutinizer-ci.com/g/ResponsiblyAI/responsibly/?branch=master

.. image:: https://img.shields.io/pypi/v/responsibly.svg :target: https://pypi.org/project/responsibly

.. image:: https://img.shields.io/github/license/ResponsiblyAI/responsibly.svg :target: https://docs.responsibly.ai/about/license.html

Toolkit for Auditing and Mitigating Bias and Fairness of Machine Learning Systems 🔎🤖🧰

Responsibly is developed for practitioners and researchers in mind, but also for learners. Therefore, it is compatible with data science and machine learning tools of trade in Python, such as Numpy, Pandas, and especially scikit-learn.

The primary goal is to be one-shop-stop for auditing bias and fairness of machine learning systems, and the secondary one is to mitigate bias and adjust fairness through algorithmic interventions. Besides, there is a particular focus on NLP models.

Responsibly consists of three sub-packages:

1. responsibly.dataset Collection of common benchmark datasets from fairness research.

2. responsibly.fairness Demographic fairness in binary classification, including metrics and algorithmic interventions.

3. responsibly.we Metrics and debiasing methods for bias (such as gender and race) in word embedding.

For fairness, Responsibly's functionality is aligned with the book `Fairness and Machine Learning

• Limitations and Opportunities https://fairmlbook.org`_ by Solon Barocas, Moritz Hardt and Arvind Narayanan.

If you would like to ask for a feature or report a bug, please open a new issue <https://github.com/ResponsiblyAI/responsibly/issues/new>_ or write us in Gitter <https://gitter.im/ResponsiblyAI/responsibly>_.

Requirements

• Python 3.6+

Installation

Install responsibly with pip:

.. code:: sh

$ pip install responsibly

or directly from the source code:

.. code:: sh

$ git clone https://github.com/ResponsiblyAI/responsibly.git $ cd responsibly $ python setup.py install

Citation

If you have used Responsibly in a scientific publication, we would appreciate citations to the following:

::

@Misc{, author = {Shlomi Hod}, title = {{Responsibly}: Toolkit for Auditing and Mitigating Bias and Fairness of Machine Learning Systems}, year = {2018--}, url = "http://docs.responsibly.ai/", note = {[Online; accessed ]} }