If you’ve typed a search phrase into Google in the past few months chances are your query was interpreted by an artificial intelligence system nicknamed RankBrain, a system that took Google engineers, search specialists and deep-learning experts almost a year to roll out.
According to Greg Corrado, senior research scientist at Google, the artificial intelligence system has been interpreting a ‘very large fraction’ of the queries typed into the search engine over the last few months, and with millions of searches taking place every second, he says the rollout went better then expected.
Google’s decision to embed the technology into every aspect of its business shows that companies really are beginning to embrace machine intelligence, with Facebook and Microsoft also using AI in some of their products. Perhaps most impressive of all is that in just a few short months, RankBrain has become the third most important signal in the Google ranking algorithm, rising above hundreds of other signals that also help determine the results we see when searching online.
Check out what Greg had to say below. (Via Bloomberg)
RankBrain makes ‘educated guesses’
Designed to handle obscure queries, RankBrain draws upon its vast amount of embedded language to make educated guesses when determining search results. Not only can it find the most appropriate results to match a query, but it’s also able to learn from mistakes to make results more accurate going forward.
What does this mean for search?
“We can always optimise for search” says MediaVision Director Thomas Schonenberger, “however the days of gaming the engines have rapidly drawn to a close. As AI technology progresses and gets more widely integrated I see our role as educating our clients in building their brand, doing the basics well and reserving that 20% of time for trying something extraordinary”
Our guess is that Google’s investment in AI tech and RankBrain will anchor Google’s position as leaders and increase their market share further. AI tech will continue to vastly improve mobile search for people who have limited mobile browsing capabilities – and improve the kind of predictive technology we already see on our mobiles every day.