Author: Brian Christian
Publisher: W. W. Norton & Company
Publication Year: 2020
Summary: The following book focuses on word2vec, an algorithm from 2013 that allows for vector computation of words. That is, word2vec allows for computations like king – man + woman = queen. Google used word2vec as part of Search and Translate. Multiple hiring platforms also used the algorithm. Researchers eventually discovered that its enormous corpus had frequent biases. For example, programmer – man + woman resulted in homemaker. Doctor – man + woman resulted in nurse. The algorithm demonstrated bias across additional protected classes as well. The story is a reminder of how important it is for data specialists to understand their data, and the risks of black box algorithms (in this case a neural network).