IBM’s Watson Group held a “Watson Cognitive Computing Innovation Day” on the web and at their Innovation Center in Foster City on Wednesday, May 21, 2014. IBM is taking Watson to market by creating an “ecosystem” that includes a developer cloud, a content store, and a talent hub.
Stephen Gold, VP of Watson Ecosystem and Market Development in the IBM Watson Group pointed out that “Cognitive Computing” differs from previous technology in that it “is not programming.” It is comprised of three main elements: natural language understanding, hypothesis generation and evaluation, and machine learning. The machine learning component enables the system to adapt and learn from training, interaction, and outcomes.
Stephen gave several examples to show how difficult natural language is. For example: ”if vegetarians eat vegetables, what do humanitarians eat?” It would be difficult to write crisp logical rules to capture all the patterns and counterexamples in language. Statistics and probability are used in the learning and IBM’s SPSS is used in Watson.
Gold said he is most excited about the effort to bring Watson to the masses using partners, including companies like Fluid. Neil Patil, President of Fluid’s Software Group showed how Fluid uses information about specific shoppers plus Watson to drive personal shopping assistants. 85 retailers want to work with Fluid. It takes them 6-9 months to deliver a SaaS application for a new retailer. The majority of the effort goes into developing the user experience, rather than the machine learning.
Sridhar Sudarsan, CTO of Watson Ecosystem, works with partners to help them become become part of and use the ecosystem. He mentioned that IBM talked with the founder of Kayak, Terry Jones, and learned that he still uses travel advisers because he likes the conversation with them (probably also because he was a travel agent early on). I assume he was hinting that Watson could supply some of the elements still missing online.
In addition to working thru partners and working on health applications, IBM is also taking Watson to market in Call Centers, in particular to provide relevant information in a timely fashion to operators.
Javier Torres gave a demo of “Watson on Watson” (WOW), a specialized instance of Watson that answers questions about Watson.
More information is available at: