Merging AI and Big Data
Posted by aonenetworks On December 7, 2015
You probably remember when IBM’s first generation super computer/AI beat Chess Master Gary Kasparov in 1997. It was a landmark achievement. Prior to that point, no AI had ever been able to beat a reigning chess champion.
Watson is IBM’s latest exploration into AI, and in every way that matters can rightly be considered Deep Blue’s “son.” He’s really impressive. Bear in mind that work began on Deep Blue in 1985, when computers were still in their infancy, and the computers of those long ago days had a mere fraction of the power that today’s machines boast. Watson takes advantage of this massive leap in capability and has been given access to Very Large Data sets which are frequently terabytes in size.
IBM has used Watson’s big, electronic brain in a number of ways, but this holiday season the company has something new for the consuming public: A shopping app called the “IBM Watson Trend App” available for free from Apple’s App Store.
Once you download and install the app, you get the results of Watson’s Big Data number crunching. While you’ve been making holiday preparations, Watson has been reading millions of comments on Facebook and other social media, in addition to a variety of discussion forums all over the web in order to collect data on what’s trending and hot this year.
As you might expect of the machine that beat two of the biggest all time Jeopardy winners in a match televised in 2011, Watson’s programming is highly complex. It combines machine learning with natural language and pattern recognition to not only gain a sense for what’s popular in human culture, but also to make predictions about what will be popular in the months ahead.
Downloading the app gives you real time updates as Watson continues to sift through the data it mines from millions of online conversations. It’s well worth a look, both for its shopping insights for this holiday season and as an exciting glimpse for what the future holds regarding AI and how we can use it to get more and better information from the Very Large Data Sets we are busily collecting.
