Fairness and Bias in Machine Learning Workshop

Jae Yeon Kim

Aniket Kesari

Renata Barreto

Avery Richards

2021-02-26

Chapter 1 Fairness and Bias in Machine Learning Workshop

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.

A gif of defendants being put into an algorithm by SELMAN DESIGN

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