"We are building into Watson a new capability to perform a dynamic psycho-linguistic profile of the user -- it understands you," said Mike Rhodin, who leads IBM's Watson group. "It can figure out why you may be asking a question so it can give you a better answer."
Also in the works are capabilities for Watson to engage in a dialog beyond its question and answer format and a reasoning facility to respond to a complex problem statement with a more detailed approach to solving it.
Rhodin described Watson as an approach to using probability theory and machine learning on natural language input, radically different to other big data algorithms such as Hadoop or Google's Map Reduce. The company is reaching out to partners large and small to bring it to a wide variety of markets.
"I tell my biz dev team to look for industries with massive unstructured data sets being produced at a rapid pace that need to be handled by experts such as law, accounting, health care and regulations -- they are great places to look at turning business models upside down," he said.
In separate steps over the past year, IBM has opened Watson to developers in startups, corporate IT departments and most recently seven universities including Berkeley and Carnegie Mellon. "Some of the great artificial intelligence universities will teach how to build apps on Watson," he said.
So far the initiative is more a center of investment than revenue. IBM announced last year it will pump a billion dollars into Watson including a $100 million venture fund. Separately, the former head of the Watson group, Manoj Saxena, started another $150 million big data fund focused on Watson
"The possibility of Bayesian answers from computers will re-configure the software stack," said Saxena in comments on stage here.
Michael Rhodin, head of IBM's Watson group, reached out to big data entrepreneurs at TIE Con.
— Rick Merritt, Silicon Valley Bureau Chief, EE Times