Fairness and Bias in Machine Learning Workshop

Bookdown https://dlab-berkeley.github.io/fairML/

DataHub

You may either download the materials from this github page and follow along in your own version of RStudio, or you can click this link to follow along in UC Berkeley's Datahub. Datahub is a campus resource that gives Berkeley affiliates access to a remote solution for working with popular programming languages like R or Python. After this workshop, you can always return to your work and even create new notebooks by visiting this link.

Overview

This workshops provides a gentle introduction to the fairness and bias in machine learning applications with a focus on the ProPublica's Analysis of the COMPAS algorithm. We revised the ProPublica's original R and Python code to increase its code readability, remixed it with other references, then published and deployed the revised notebook using bookdown and GitHub page.

Outline

1. Bias in the data

• Risk of Recidivism Data
• Risk of Violent Recidivism Data

2. Bias in the algorithm

References

For more information on the ProPublica's Machine Bias project, we encourage to check out the following references.

• Argument by Julia Angwin, Jeff Larson, Surya Mattu and Lauren Kirchner

• Counterargument by Sam Corbett-Davies, Emma Pierson, Avi Feller and Sharad Goel

• Methodology