Google's AI: RankBrain and TensorFlow
Google has been investing heavily in Artificial Intelligence for videos, speech, translation and, recently, search.
For the past few months, a very large fraction of the millions of queries a second that people type into the company’s search engine have been interpreted by an artificial intelligence system, nicknamed RankBrain, writes Bloomberg in an article reporting on the new AI. About 15 percent of searches sent daily haven't been seen by Google before, and RankBrain is reportedly useful in answering those queries.
RankBrain uses artificial intelligence to embed vast amounts of written language into mathematical entities, called vectors, that a computer can understand. Once RankBrain analyzes the text through vectors, it can isolate words or phrases it doesn't understand. It can then guess the meaning based on similar words and phrases and filter the results accordingly.
Google uses hundreds of 'signals' designed to bring users the most useful search results, but they’re all based on discoveries and insights that people in information retrieval have had, and there’s no learning involved. According to Google, RankBrain has already turned out to be very effective, becoming the third-most important signal contributing to results. “Machine learning is a core transformative way by which we are rethinking everything we are doing,” said Google’s Chief Executive Officer Sundar Pichai.
RankBrain is not the only way Google has been working on AI software. Recently, the tech giant released TensorFlow, a new AI system used for everything from recognizing speech in a crowded room to finding photos of your cat, and, as an example, for the recognition of foreign words on a street sign to translate them in real-time. What's even more interesting about it, is that Google makes the software open source, so that anyone can access and edit the code.
According to Google this software is already five times faster than the first generation, and can be run on thousands of computers or on a single smartphone. The reason for Google to make TensorFlow open source is to stimulate innovation and to make it easier for researchers to share their ideas and code. Google has released a version of this new technology running on a machine, a version that acts on multiple machines will follow in the near future.
"Machine learning is still in its infancy—computers today still can’t do what a 4-year-old can do effortlessly, like knowing the name of a dinosaur after seeing only a couple examples, or understanding that “I saw the Grand Canyon flying to Chicago” doesn’t mean the canyon is hurtling over the city. We have a lot of work ahead of us. But with TensorFlow we’ve got a good start, and we can all be in it together", says Google.
Watch Google's presentation of TensorFlow: