Skip to main content

Computing – Patterns and predictions

ORNL's open-source software mines for insights in Big Data, enabling timely detection of useful information such as fraud within a healthcare service provider network.

January 5, 2016 – Drawing connections between seemingly disparate and vast amounts of text could become easier thanks to software developed by Oak Ridge National Laboratory. Researchers Matt Sangkeun Lee and Sreenivas Rangan Sukumar bring “six degrees of separation” to the computational field with holistic graph analysis technology, a smart data tool that scales on Cray’s URIKA-GD. The open-source software generates webs of patterns and predictions from stored data while automatically extracting tidbits of useful knowledge. For a world of disparate data such as social media or healthcare, this is a promising tool for visualizing and discovering new information from stored data. “We have the hardware and the software,” Lee said. “It’s not just theory anymore.”