Devs Will Get to Put IBM's Watson Through Its Paces
The Watson system is "extremely good at coalescing, ordering or querying large volumes of complex information, so its use in areas like healthcare where it can act as a diagnostic consultant is a natural fit," said Pund-IT's Charles King. Watson will allow IBM "to develop or deliver cognitive services that other vendors will be hard-pressed to match, let alone surpass."
11/14/13 3:05 PM PT
IBM on Thursday announced it will open up its Watson cognitive platform, allowing software developers to access it in the cloud starting next year.
The technology is based on its Watson supercomputer.
Participants will build apps incorporating Watson's cognitive computing intelligence.
A cognitive computing system is "capable of continually learning and refining itself through human/computer interaction," Rob High, CTO of IBM Watson Solutions, told TechNewsWorld.
"By creating applications that learn from experience, improve with each interaction and outcome, and answer users' complex questions with data-driven insights, IBM's ecosystem partners will gain a competitive advantage and deliver solutions to needs in the marketplace that have not been met," he asserted.
Full-blown Watson solutions "will be out of the price range for the vast majority of organizations ... so leveraging state-of-the-art Watson technologies in a cloud setting makes great sense for both developers and end customers," Charles King, principal analyst at Pund-IT, pointed out.
IBM's Plans for Watson
App developers ranging from startups and emerging VC-backed companies to established software firms will be able to play.
They will get resources -- including a developer toolkit, educational materials for self-service training, and access to Watson's application programming interface -- from the Watson Developers Cloud.
They will be able to test their cognitive apps on the Watson Developers Cloud as well. The developers can either use their own data or access the IBM Watson Content Store, which will offer third-party content.
IBM will connect developers with skilled professionals who can help ready their products and services for the market. The company has committed to the program more than 500 of its subject matter experts in areas such as design, development and research.
Training the Talent
Big Blue plans to offer a talent certification program through various Talent Partner participants, beginning in 2014.
Whether developers will have to pay for certification will depend on the talent providers that sign up for the ecosystem, IBM's High said.
"IBM will be building certification standards and making those standards available to talent providers to administer," he continued.
Program Road Map
Developers first must identify and develop use cases that show how embedding a Watson capability into an app will offer a unique value proposition.
They must then identify content owned or licensed through the Watson Content Store that will support their app.
Next, they will design and develop the app using a sandbox instance of Watson in the Watson Developer Cloud.
That will be followed by training and testing the offering through the Watson Experience Manager's training and user experience configuration tools.
Finally, the partners will deploy their apps through IBM's hosting solutions or over their own infrastructure.
The apps can run on any OS developers select, IBM's High said.
Breaking New Ground
Organizations that have the rights to use or license general and domain-specific content -- such as publishers, researchers and possibly social forums -- are ideal candidates for IBM's Watson Ecosystem program.
"We have five initial partners -- Fluid, MD Buyline, Welltok, Elance and Healthline -- who are key examples of the types of organizations we are welcoming into this cloud-based program," IBM's High said.
The Watson system is "extremely good at coalescing, ordering or querying large volumes of complex information, so its use in areas like healthcare where it can act as a diagnostic consultant is a natural fit," Pund-IT's King told TechNewsWorld.
"In Watson [IBM has] a platform which should allow them to develop or deliver cognitive services that other vendors will be hard-pressed to match, let alone surpass," he observed.
IBM has been seeking ways to monetize Watson since the supercomputer's success in the Jeopardy! TV game back in 2011.
"Watson ties back into IBM hardware, software and services," Rob Enderle, principal analyst at the Enderle Group, told TechNewsWorld.
"It has a long tail of revenue … but this is potentially not only a very lucrative offering, it has massive value to the customer," Enderle continued. "This could easily be the most profitable solution IBM has ever had and still be underpriced."