Tutorials

The tutorials complement the main content of this book by diving deeper into some of the topics via case studies.

Measuring Group Fairness

Explore how to measure notions of group fairness via a disaggregated analysis in Python using Fairlearn. As a running example, we consider pre-trial risk assessment scores produced by the COMPAS recidivism risk assessment tool.

---

After this tutorial you are able to:

• perform a disaggregated analysis in Python using Fairlearn;

• describe the relevance of reliability of estimates and validity of measurements in relation to quantitative fairness assessments;

• explain the incompatibility between three group fairness criteria;

Fairness-aware Machine Learning

Explore several fairness-aware machine learning techniques in Python using Fairlearn. As a running example, we consider pre-trial risk assessment scores produced by the COMPAS recidivism risk assessment tool.

---

After this tutorial you are able to:

• apply techniques for fairness-aware classification in Python

• examine strengths and limitations of fairness-aware classification techniques