to automate tasks, a popular data-backed model called word embedding has also picked up entrenched social biases. Earlier this year, for instance, gender bias examples started cropping up on social media with Google Translate. Word embedding works by linking words to a vector of numbers, which algorithms can the use to calculate probability. A 2016 study that trained word embeddings models on articles on Google showed gender stereotypes “to a disturbing extent,” according to its researchers. McCurdy says that there isn’t anything necessarily wrong with the word embedding model itself, but it needs human guidance and oversight.

Source: Forbes February 15, 2018 18:22 UTC