Sunday 11:15–12:00 in D105 Audimax

Biases are bugs: algorithm fairness and machine learning ethics

Françoise Provencher

Audience level:
Novice

Description

Biases are bugs. They need to be found, fixed, and learnt from. A mix of good ethics and good engineering practices can get us a long way towards that goal.

In this talk you'll learn what biases are, what software tools can help, and how to adopt engineering practices that can make your algorithms fairer.

Abstract

Algorithms can make decisions, and these decisions can have an impact on people's lives. By feeding data into these algorithms, they can reproduce or amplify our societal biases and take unfair decisions.

Biases are bugs. They need to be found, fixed, and learnt from. A mix of good ethics and good engineering practices can get us a long way towards that goal.

In this talk, you will learn what biases are, see examples of algorithms gone wrong, and explore some software tools you can use and engineering practices you can adopt in your own work to make your algorithms more fair.

Subscribe to Receive PyData Updates

Subscribe

Tickets

Get Now