Doodling AI : Man is to Computer Programmer as Woman is to Homemaker?

Second post in a series where I doodle a gentle overview of ML/AI research papers that I find interesting. If you find it interesting, please refer to the paper for more details.

Today’s topic: Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings (2016)

This is a very important paper that provided tools to quantify bias in word embeddings and algorithm to debias them. However, it handles biases one at a time, requires a list of paired words as input (eg. he-she) and doesn’t specify how to handle multiple categories (eg. he-she-they).

Thanks for reading. Eager to hear your thoughts in comments!