With all the money being spent on, and all the futuristic talk about about big data, machine learning, artificial intelligence and all things in between, it’s easy to forget that Microsoft and Google — two of the companies leading research in these technologies — still have large businesses in web search. So as cool and potentially life-altering as AI might be in fields such as medicine, we’ll probably continue to see the signs of things to come in search engines first.
It’s big business and a great testing ground.
Take, for example, Microsoft Bing’s new predictions feature that tries to predict the outcomes of popular fan-voting show such as The Voice, American Idol and Dancing with the Stars. Bing does this by analyzing a number of signals, including searches, Twitter and Facebook data, and, presumably the outcomes of previous episodes. (Hat tip to Search Engine Land for spotting this Bing blog post, as well as the following one.)
Microsoft is also boasting about how Bing’s image search — which includes filters for all sorts of variables — is superior to Google’s image search. A blog post explaining the differences reads like an homage to advances in computer vision and object recognition (and, very possibly, deep learning) without ever mentioning those terms. Google, of course, has been using these technologies to power new image search features, too.
Might improved machine learning someday give us better teleconferencing, better office software and a Terminator-like view of the world around us, or help doctors better analyze cancer cells? Sure. But first it will make sure we don’t waste our time looking at poor-quality celebrity pics or betting on reality show losers.
For more on the importance of machine learning and AI research at companies like Microsoft, check out this talk from Stucture Data with Microsoft Research machine learning manager John Platt.
[protected-iframe id=”5b1d4c689eacca6ae02657d0c2f287a5-14960843-6578147″ info=”http://new.livestream.com/accounts/74987/events/2795568/videos/45661245/player?autoPlay=false&height=360&mute=false&width=640″ width=”640″ height=”360″ frameborder=”0″ scrolling=”no”]
