Google will answer your questions instead of providing search results

Ask Google “Who’s the highest paid sportsman” on its search app and it’ll give you the right answer instead of search results. Also, it adds a video of top paid athletes in the world.

Google answers these questions with the help from deep neural networks, a type of artificial intelligence quickly changing not only Google's search engine but the whole company and, well, alternate giants of the internet, from Facebook to Microsoft. Deep impartial nets are pattern recognition systems that can figure out how to perform particular tasks by breaking down unfathomable amounts of data. For this situation, they've figured out how to take a long sentence or paragraph from a significant page on the web and extract the upshot—the information you're searching for.

These sentence compression algorithms just went live on the desktop incarnation of the search engine. They handle a task that is really basic for humans however has generally been very troublesome for machines. They demonstrate how deep learning is propelling the craft of natural language understanding, the capacity to comprehend and react to natural human speech.

“You need to use neural networks—or at least that is the only way we have found to do it,” Google research product manager David Orr says of the company’s sentence compression work. “We have to use all of the most advanced technology we have.”

To train Google's artificial Q&A brain, Orr and company likewise utilize old news stories, where machines begin to perceive how features serve as short synopses of the more drawn out articles that take after. In any case, for the time being, the company still needs its group of PhD linguists. They exhibit sentence compression, as well as really label parts of speech in ways that help neural nets see how human language works.

This sort of human-helped AI is called "supervised learning," and today, it's exactly how neural networks work.