Using ORNL’s Summit supercomputer, scientists can comb through millions of medical journal articles looking for possible connections among FDA-approved drug therapies and known COVID-19 symptoms. Credit: Dasha Herrmannova/Oak Ridge National Laboratory, U.S. Dept. of Energy
Scientists have tapped the immense power of the Summit supercomputer at Oak Ridge National Laboratory to comb through millions of medical journal articles to identify potential vaccines, drugs, and effective measures that could suppress or stop the spread of COVID-19.
A team comprising researchers from ORNL and Georgia Tech are using artificial intelligence methods designed to unearth relevant information from about 18 million available research documents. They looked for connections among 84 billion concepts and cross-referenced keywords associated with COVID-19 — such as high fever, dry cough, and shortness of breath — with existing medical solutions.
“Our goal is to assist doctors’ and researchers’ ability to identify information about drug therapies that are already approved by the U.S. Federal Drug Administration,” said ORNL’s Ramakrishnan “Ramki” Kannan.
A massive subset of 6 million documents dated between 2010 and 2015 took 80 minutes, and the entire 18 million will take less than a day to run on Summit. Results will be shared with medical researchers for feedback, which will inform adjustments to improve future calculations.
Be the first to comment on "Artificial Intelligence Equipped Supercomputer Mining for COVID-19 Connections in 18 Million Research Documents"