The mighty RankBrain is an algorithm learning artificial intelligence system, the use of which was admitted by Google in 2015. It helps Google to process search results and provide more relevant search results for users. In a 2015 interview, Google said that RankBrain was the third most important factor in the ranking algorithm along side with links and content.
If RankBrain sees a word or phrase it isn’t familiar with, the machine can make a guess as to what words or phrases might have a similar meaning and then filter the result accordingly, making it more effective at handling never-before-seen search queries or keywords.
Search queries are ordered into word vectors, also known as “distributed representations,” which are very close to each other in terms of linguistic similarity.
RankBrain then tries to map this query into words (entities) or clusters of words that have the best chance of matching it.
RankBrain attempts to predict what people mean and records the results, which adapts the results to provide a superior user experience. It actually works too, Well!
RankBrain is considered the 3rd most important signal in determining ranking on Google (aka SERPS). Google has not admitted to any order of importance of factors, only that Rankbrain is one of the three most important of its search ranking signals. When it is offline, RankBrain is fed with batches of past searches and it then learns by matching search results together. This makes Rankbrain have a better performance as it ages and learns.
Google has admitted that it uses tensor processing unit (TPU) ASICs for actually processing RankBrain requests.