Bias Matters! What’s Fairlearn, and why should I care?

Original article was published by Bruno Cordeiro on Artificial Intelligence on Medium

The trouble with bias

The concern about bias in A.I. is not something new. Kate Crawford has risen her concerns back in 2017, at NIPS 2017 Keynote. If you want to watch the full video, you can access The trouble with bias.

Part of the problem is that people trained as data scientists who build models and work with data aren’t well connected to civil rights advocates a lot of the time.
– Aaron Rieke

Machine learning engineers must pay attention when developing a model to avoid replicate human processes that were historically a source of bias.

A 2012 paper, studied the influence of demographics on the performance of face recognition algorithms and found they were less accurate when identifying the faces of black people, along with women and adults under 30.

Avoiding bias requires an understanding of both very complex technology and very complex social issues, which leads to an understanding of techno-social systems.

Techno-social systems refer to the circumstance that the web cannot be defined without connection to the human social realm.
– Celina Raffl