Google RankBrain AI Tasked With Handling Your Most Outlandish Search Requests
As we wind down 2015, it's not hard to see some trends arise when looking back at the months leading us up to this point. Right from the get-go, an obvious trend with the biggest companies out there has been the adoption of machine-learning techniques. While machine-learning isn't new, it's now being treated as something of huge importance, and it's definitely going to be a big part of our future.
Late last year, we learned that Microsoft had adopted NVIDIA graphics cards to help it accelerate its machine-learning capabilities, and this past May, the company released a neat age-guessing tool. While that's more for fun, machine-learning can be used for more important things as well, such as delivering relevant information to you, fast.
Well, it should come as no surprise that Google is one company that's taking full advantage of machine-learning capabilities, and search is at the forefront of its initiative. In a recent Bloomberg report, we learn of one of Google's latest "signals" used to help improve search results. Called RankBrain, this signal takes advantage of machine-learning to help interpret queries that simply would have been considered too complex years ago. An example given is, "What’s the title of the consumer at the highest level of a food chain?"
While RankBrain is just one of "hundreds" of signals Google uses to improve its search results, it's proving to be one of the most important. In fact, senior research scientist at Google Greg Corrado said that if it were disabled, it'd be doing a major disservice to Google's users. He even compared it to being equivalent to disabling half of the pages on Wikipedia. Now there's a grand statement.
Currently, RankBrain handles about 15% of search queries, but given its impact up to this point, and its ever-evolving mechanic, we could see that share rise in the future.