RankBrain is an artificial intelligence system by Google, which was implemented in main algorithms of Google search engine in 2015. Its main task is to interpret queries entered by users in a way different from the one used so far, and to translate them to various possibilities to formulate the same phrase.
As a result of this process, RankBrain algorithm is able to “guess” what a given user of the search engine is looking for, match similar phrases, and finally display the most accurate ones and personalise output data. Thus, a person using the search engine has an impression that Google guesses and predicts their intentions.
The essence of the algorithm
RankBrain is also a perfect tool for analysing new phrases which are entered into the search engine for the first time and does not need to catalogue them in detail in the data base before this process. It is able to do so during the process. According to the management of the Silicon Valley Giant, RankBrain is the third most important algorithm responsible for the order in which search results in Google are displayed.
How does it work?
From a mathematical perspective, according to Greg Corrado — senior search specialist — RankBrain translates written text into vectors which the computer can understand. Thus, the computer is able to “understand” phrases unknown to the search engine before, “guess” their meaning by looking for words with similar meanings, and display results more accurate than previously — before it was introduced. Since 2015, chances have been very slim for Google to “tell us” that it has not been able to find a single answer. To prove that, Greg Corrado says that RankBrain helps Mountain View to recognise 15% more queries than in the years before its introduction. The objective of Google is to maintain its advantage on the market of search engines and to make artificial intelligence capable of dealing also with ambiguous and more complex queries.
RankBrain, combined with the whole range of AI algorithms functioning for Google, especially with the voice recognition system, may cause that users have an impression that they are speaking with a live human being providing them with reliable and exhaustive information.