Google Offers Cloud-Based Learning Engine

There are already many companies that are using machine-learning algorithms. These algorithms enable software to learn how to respond to new information or events with intelligence. For example, Amazon has product recommendation. The software chooses products related to ones that you have clicked on or purchased and recommends that you look at them. It chooses products that you may be interested in and may want to buy. Pandora also uses an algorithm that can guess what other artists you may want to listen to. The service that Google has recently launched will bring this “learning” ability to many more apps. For example, a Google-Hosted email app could sort out your junk mail from important mail by “remembering” the senders of the mail that you don’t open and the mail that you always open. Instead of developers coming up with their own algorithm, Google makes their Google Prediction API available.
The way that Google’s service works is with a “kind of machine-learning black box -- data goes in one end, and predictions come out the other.” “Developers can deploy it on their site or app within 20 minutes,” says Green. This will put bring the predictive API into more apps easily making it a smother and quicker transition into this new technology. “We're trying to provide a really easy service that doesn't require them to spend month after month trying different algorithms.” Only large companies like Amazon have been able to make and use prediction API but now that Google will make it available to more companies, it will have a leveling effect. This will help other companies to get a better competitive edge.
Some people are already making plans on what they want to do when they gain access to the Prediction API. They could use it in applications used in popular websites such as Facebook and Twitter. Chris Bates agrees that Google's black box will enable wider use of machine learning. “Today it is good at predicting which language text is in and also sentiment analysis, for example to pick out positive and negative reviews,” he says. Even before the API is released, people are coming up with great ways to use it to improve applications.
New creations such as the Prediction API help take technology one step further. These new programs make using a computer even easier and take less steps. More things are becoming automated so that we don’t have to do as much work. Applications can predict what we want to buy, listen to, type, or the next person we want to add as a friend on Facebook. I think that this API will allow programmers to easily put this cool predictive feature into most of the online applications and websites that we use. And since the API is in the clouds, not requiring users to download anything, it will run much smoother. Anyone with internet will be able to experience it. Most people might not even notice when more websites start sing it because it will seem so seamless and natural.

http://www.technologyreview.com/computing/26093/