Challenge: Fairness through Unawareness

Implementing fairness through unawareness

In this lesson, we will implement the simplest possible mitigation algorithm: fairness through unawareness. We found that the model can discriminate because of gender. Why don’t we remove this feature? If the model doesn’t know something, it cannot be unfair because of it. Right? Not quite. The reality is, as always, more complicated. But for now, let’s focus on implementing the solution. We will get back to the question in a moment.

This notebook contains a simple exercise. Answers can be checked below in the Jupyter Notebook.