Watson has big plans for the future

The supercomputer's possibilities are seemingly endless
By Mike Miliard
01:00 AM

It's been a busy three years for IBM's Watson supercomputer, ever since it first won a whopping $35,734 on Jeopardy! in 2011, crushing all comers on the TV quiz show.

In just the past six months or so, Watson has joined the fight against cancer, putting its cognitive computing capabilities to work at Texas MD Anderson Cancer Center, and has rolled out as a development platform for application creation, powering healthcare apps by the likes of MD Buyline and Welltok, with many more to come.

Heck, it even tried its hand at cooking for the first time this past March, helping develop the menu for a food truck at the SXSW festival, coming up with creative dishes by coming through and combining ingredients from a list of thousands of recipes.

"Its as if it was digesting a giant cookbook," said Steve Abrams, IBM's director of Watson Life at IBM, in a press statement. "From reading that cookbook it has learnt an awful lot about different ingredients that are often used in different cuisines."

Watson has learned an awful lot about our world in general – and healthcare in particular – these past few years. Clearly, this "thinking" machine is getting closer to world domination every day.

At HIMSS14 in Orlando, Fla., earlier this year, Stephen Gold, IBM's vice president of Watson Solutions, sat down with Healthcare IT News to take stock of where Watson has been – and look ahead to new capabilities perhaps not yet dreamed of.

"In 2011, two things happened for us," said Gold. "One was understanding whether or not we could actually deliver an experience that had never been done before. The whole Jeopardy! phenomenon, outside of the novelty, was really the test of a whole brand new style of computing."

Second, was the task of "figuring out what commercial applications this had," he said. "There was some uncertainty early on about where to take this commercially, because we didn't really know how it was going to perform. Healthcare, which was where we chose to begin, probably wasn't the best choice – we didn't pick the easiest industry. But it was a great proving ground."

And how. After wowing Alex Trebek, Watson quickly moved on to deployments of various types with WellPoint, where it did TK, and at giants like Memorial Sloan Kettering and Cleveland Clinic, where it did TK.

Those projects led to an "understanding that the technology was actually fairly consistent across those applications – which wasn't the original assumption," said Gold.  
"In 2011 we assumed everything would be bespoke, that we'd have to tailor it to the specific use case."

Instead, "What we found was that the content changes, so the corpus that we use is different, and the user interface is different. A doctor experiences it through an EMR; a patient would never experience it through an EMR. And yet, the core technology didn't change. That was sort of an 'a-ha' moment that allowed us to move forward."

Another evolution has been the Watson development platform itself. "Everything we've built since probably early last year has been on top of a single platform," says Gold. "You talk about this ecosystem, this idea of working with partners, so they can incorporate a Watson experience, it's all on top of that one platform. So that gives us true scalability to the technology."

This past January, IBM launched its Watson Group, which seeks to accelerate the evolution of new species cognitive computing tools. Among the first new technologies developed by the group and delivered via the cloud are Watson Discovery Advisor, which focuses on pharmaceutical and biotech R&D.

Watson Analytics, meanwhile, sets its sights on big data, able to delve into questions posed in natural language; Watson Explorer takes a broader mandate, seeking to help organizations launch data-driven initiatives – and get at insights from them – more easily.

Want to get more stories like this one? Get daily news updates from Healthcare IT News.
Your subscription has been saved.
Something went wrong. Please try again.