Search
Homepage This page's url is: -crn- Rankings and Research Companies CRNtv Events WOTC Jobs Tech Provider Zone

IBM Introduces First of Several Watson Data Kits To Jumpstart AI Projects

Curated data sets and pre-trained algorithms are meant to launch enterprises over the initial, and often most-burdensome, obstacle in deploying intelligent apps.

To accelerate industry adoption of artificial intelligence, IBM introduced on Tuesday the first of several Watson Data Kits intended to get partners and customers out of the gate faster in developing intelligent apps.

The kits offer curated data sets and pre-trained algorithms that remove a bottleneck often dominating roughly 80 percent of a project lifecycle – the initial phase of gathering, cleansing and ingesting data to train machine learning models, said Kourosh Karimkhany, offering manager at Watson Content who previously headed Weather Underground, a subsidiary of IBM's Weather Company business.

"We found the toughest part of getting started with Watson, and AI in general, is it just takes way too long, that initial phase," Karimkhany told CRN. "It is a huge, huge pain point."

[Related: Q&A: IBM's Watson Leader On 'Dark Data' And The Age Of Reason]

For typical projects, Karimkhany said, several data scientists work on mundane data preparation tasks for months at a time before the machine learning process even begins.

"That's a point of inertia. We think having these data kits will overcome that inertia. We can take the initial phase from months to minutes," he told CRN.

The first kit demonstrating the concept focuses on powering intelligent apps for the travel industry. United Airlines is an initial partner.

IBM culled information from internet sources and partners on some 20,000 cities around the world, with 300,000 points of interest that have been categorized and tagged. As more information comes in from users, those models will continue to evolve.

Future kits will be more in IBM's traditional wheelhouse – solutions for retail, health care, insurance, banking, and other highly regulated industries. Those data sets will facilitate the development of base models for enterprise clients to use to build customer-care applications quickly.

"Over the coming months as we look at more use cases, we're going to get a better sense of what the specific pain points are for each of the industries," Karimkhany told CRN.

IBM's channel has yet to be engaged on the concept. But with the launch of the first Watson Data Kit, Big Blue plans to start talking to prospective business partners about expanding their emerging AI practices around those solutions.

Back to Top

Video

 

sponsored resources