BERT

Contrary to belief, BERT is not an algorithm update. Instead, it is a machine learning model that updates Google’s processes to improve Natural Language Processing. It is designed to improve how the search engine processes or understands the queries users search for in relation to all of the other words used within the search query.

Before BERT, Google used a word-per-word process, so the update’s introduction means that it works much more intelligently.

BERT is designed to understand words, phrases, and entire content just as we do, including any ambiguous language. It is particularly important thanks to the advent of voice search using devices such as Siri and Alexa.