IBM Uses RESTful APIs to Turn Watson into a Cloud Service

The combination of natural language processing and advanced text analytics is giving rise to a new class of cognitive applications that have the potential to radically transform the way entire industries operate. The most famous instance of a cognitive application is, of course, IBM Watson, the supercomputer that IBM built on top of Power processors to best the champions from the Jeopardy quiz show.

Beyond playing games, however, the ability to easily query a knowledge base of expertise that gets smarter with each successive correct answer it generates has the potential to put a massive amount of expertise directly into the hands of the average. Now IBM is moving to put that power in the hands of developed with the launch of IBM Watson Developer Cloud, which includes a software development kit that will allow developers to build applications on top of IBM Power systems running on an IBM cloud while at the same time exposing a set of RESTful APIs that will make it possible to invoke those applications from within another application. In addition, IBM plans to make third-party applications developed on this Platform available to customers through a new IBM Watson Content Store.

The first three independent software vendors to participate in the program include Fluid, which Builds on-line shopping applications for retailers; MD Buyline, a provider of supply-chain applications for hospitals; and Welltok, which built a social health application for end users. All three are scheduled to have applications available on the IBM Watson cloud in early 2014.

IBM is also partnering with providers of sources of data that can be consumed by Watson. They include Healthline, a provider of health information services; and Elance, an online marketplace for freelancers.

According to Rob High, a chief technology officer and IBM fellow in the IBM Software group, IBM plans to work with several venture capital firm to help promote the development of both cognitive computing applications and additional sources of data services that could be consumed by Watson.

Just like any Web application, cognitive applications will likely be most useful when they are invoked within the context of a business process. For example, a healthcare provider recording a patient's symptoms could dynamically query a cognitive application running in the cloud to determine the best course of treatment to pursue. Similarly, a law firm would be able to make use of a cognitive application to determine what laws best apply to any given case.

Today, people spend countless hours researching information that an application that supports natural language processing could instantly put at their fingertips or a voice command away. To enable that IBM wants to build an ecosystems of Watson applications in the cloud that could be dynamically invoked by thousands of other applications, High says.

The challenge with creating cognitive applications is that right now most developers don’t have access to supercomputers based on IBM Power systems. By making Watson available as a service in the cloud IBM is removing hardware issues as a barrier to the development of next-generation cognitive computing applications. To accomplish that commercial success, High says, IBM has almost completely revamped the Watson system that appeared on Jeopardy!

No doubt other ecosystems of cognitive applications will emerge in time. But for developers looking to create truly unique applications that have the potential to transform the economics of entire industries, IBM Watson in the cloud provides an opportunity to get started building these applications without having to spend millions of dollars just to find out whether they will work or not.

Be sure to read the next News Services article: Deloitte Unfurls CloudMix Integration Service