Case Study 4: Facial Recognition

Task

Read this article on the use of facial recognition in machine learning.

Recap

The data we use to train our model affects the results of our model. It’s possible for our model to be biased towards/against certain groups, simply from having a group under/over represented in our dataset. Thus, it is important to evaluate the diversity in your dataset and evaluate our results on a diverse dataset and to keep in mind what we are using this model for.

Further reading:

• See this talk by Joy Buolamwini on how her research helps fight bias in algorithms.