RankBrain: Google’s Machine Learning AI

Google announced its new machine-learning artificial intelligence system called RankBrain in October. RankBrain is one of the hundreds of signals Google uses to deliver the best search results. The difference between RankBrain and other signals is it can actually interpret search queries. This is real AI in practice RankBrain allows computers to think for themselves and actually understand language.

Search queries are becoming more ambiguous and colloquial. We ask Googles search engines questions like we would speak to another person. The result is that every day there are brand new queries that the search engine has never seen before. RankBrain was created to be able to generalize and know that a search sounds like something it has seen in the past, even though it has never seen it before. This new machine learning will be able to deliver better search results and is becoming one of the most important signals in all the hundreds of signals in Googles search engine.

If you want to get into the nitty-gritty of it, RankBrain interprets written words into mathematical entities called vectors. If RankBrain sees a word/phrase/vector that it is not familiar with, it can make an educated guess based off of similar vectors. Google has published an open source toolkit called word2vec with the intention of learning the meaning behind words. To learn more about word2vec, read Googles blog post here.

Greg Corrado, a senior research scientist at Google, said, In the few months it has been deployed, RankBrain has become the third-most important signal contributing to the result of a search query. So although RankBrain is going to become a very important factor in a large fraction of our searches, there are still other important signals to the algorithm. Some other important signals include links, trust, social reputation/shares, content, architecture, and html. I believe the three most important signals that Corrado was referring to include links, content, and now RankBrain.

RankBrain is a huge step in AI and machine-learning. With the integration of RankBrain in Googles search algorithm, we will be able to interact with computers more easily. We will be able to use natural language when interacting with computers, get search results more efficiently, and move on with our day.

Leave a comment

Your email address will not be published.