Since 1986 - Covering the Fastest Computers in the World and the People Who Run Them

Language Flags
October 18, 2013

IBM Research Rolls Out Accelerated Discovery Lab

Tiffany Trader

Last week, IBM opened up a new competency center called Accelerated Discovery Lab – ADLab for short – based out of the IBM Research – Almaden site in San Jose, Calif. In this world of sensors and clouds and ubiquitous connectivity, researchers at ADLab are working to transform data into meaningful intelligence for a broad range of industries.

“The lab, which has physical and virtual components to it, is a cross-Research effort to create an environment that can accelerate our ability to get insight or predictive capability out of data, leveraging a rich set of hardware, software and domain experts,” notes a company statement.

ADLab is not starting from scratch. The research center will apply the smart analytics software that is the basis for IBM’s Watson question-and-answer (Q&A) technology to a number of projects in the accelerated discovery space. Initial focus areas include water management, social media analytics and drug discovery.

What’s different here – indeed what’s game changing about big data analytics – is that the researchers are working in partnership with machines to mine data for gold, but they don’t know what that gold is yet. It’s machine-assisted brainstorming.

In an interview with EETimes, Jeff Welser, director of strategy and program development at IBM, drove home this point. “The Accelerated Discovery Lab shifts the focus away from just looking through big-data question-and-answer style,” Welser said. “Instead of looking for answers that are already known – where its just a matter of finding them – we are learning how do search for things that are not yet known.”

In this video, IBM Research – Almaden lab director Michael Karasick takes viewers on a behind the scenes tour of the new lab and discusses some of the work going on there. Karasick describes the effort as part of a C-change “to be a little less compute-centric and a little more data-centric.”