On Sunday, October 25, 2015, Bloomberg’s Jack Clark broke some serious news regarding how Google processes search queries. If you’re familiar with SEO, you know that any insight Google gives into its inner workings is as priceless as it is rare. But after posting better than projected earnings, Google’s executives were in the mood to spread some love.
After dissecting the article and the insight from Greg Corrado, a senior research scientist at Google, here’s what we know about the company’s artificial intelligence advancements:
1. What RankBrain actually is.
RankBrain is the first glimpse into Google’s machine learning and artificial intelligence capabilities. After the written input is converted into numerical data, the computer can identify those strings of which it is unfamiliar with and hypothesize other strings that share similar characteristics.
“RankBrain uses artificial intelligence to embed vast amounts of written language into mathematical entities — called vectors — that the computer can understand. 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 filter the result accordingly.”
The key to RankBrain (and all artificial intelligence) is the ability to learn. This signal can identify patterns and insights once only detectable by people.
2. RankBrain is handling a large portion of Google’s search queries.
Google sees approximately 3.5 billion search queries every single day. Would you believe that 15 percent of those, or 525 million queries, it has never encountered before? RankBrain is a large contributor to processing those 525 million queries. And it’s doing a darn good job. The company found that turning RankBrain off “would be as damaging to users as forgetting to serve half the pages on Wikipedia.”
3. RankBrain outperforms Google’s search engineers.
Further proof of RankBrain’s efficacy lies in its ability to synthesize queries with a higher rate of accuracy than its human predecessors. Google revealed that in comparative testing of query ranking, Google search engineers were correct 70 percent of the time, while RankBrain enjoyed a success rate of 80 percent.
4. RankBrain is the 3rd most important signal accounted for within Hummingbird.
We know that the algorithm Google uses to process search queries is called Hummingbird, which understands both the context and the content (the keywords) of a conversational query such as this, meaning, the words “get people to” is equal to the keyword “traffic.”
Until Sunday, all that was known about Hummingbird’s inner signaling factors was that there were ‘hundreds’ of them. While we haven’t gotten much more insight than that, Corrado did reveal an interesting piece of information about the signals within Hummingbird as a whole:
“In the few months it has been deployed, RankBrain has become the third-most important signal contributing to the result of a search query.”
5. ‘AI Wars’ are upon us.
Facebook and Microsoft (i.e. Bing) are already using artificial intelligence to personalize and broaden the capabilities of their search engines. Now that we’ve seen Google’s massive gains at the foot of AI capabilities, it’s only a matter of time before greater strides in AI are brought to the forefront.