March 27, 2020

COVID-19 Vaccine Developers Gain Enhanced Access to Supercomputers

Federal agencies, academic institutions, and industrial partners are joining forces to combat COVID-19 using artificial intelligence.

Earlier this week, the White House announced the creation of the COVID-19 High Performance Computing Consortium, a coalition of federal agencies, academic institutions, and partners in industry to provide researchers working on vaccine development access to supercomputers, with an aim to expedite the end of the current pandemic.

There are currently three research institutions in the consortium: MIT, the University of California, San Diego, and Rensselaer Polytechnic Institute in New York. They are joined by federal agencies including the National Science Foundation, NASA, and laboratories within the Department of Energy along with industry juggernauts IBM, Google Cloud, Microsoft, Amazon Web Services, and Hewlett Packard.

The announcement of this consortium comes on the coattails of the release of the COVID-19 Open Research Database, an open-access project from the National Institutes of Health that provides data from more than 44,000 articles about SARS-CoV-2 available for AI analysis.

“Something we’ve seen as this pandemic has taken hold is the entire research community is ready, willing and eager to help assist the American people in this time of crisis,” US Chief Technology Officer Michael Kratsios tells The Wall Street Journal.

The 16 supercomputers in the network represent 330 petaflops, 775,000 CPU cores, and 34,000 graphics processing units. As a comparison, it would take nearly 100,000 new MacBook Pros working at peak performance to match the results of the supercomputers. Researchers can apply for access to the computers to run AI-assisted simulations of specific molecules related to vaccine development and new therapies. The simulations can help predict mutations and evolution of the virus, which are typically difficult for ordinary computers because of the number of variables that need to be considered.

Another supercomputer collective that is taking on COVID-19 is the C3.ai Digital Transformation Institute(DTI), uniting the eponymous artificial intelligence company C3.ai with various academic institutions and private companies. The institute is planning on investing more than $365 million into grants during the first five years of operation to advance AI techniques for use in the public and private sector, but in these early days, the focus will be on containing COVID-19.

C3.ai DTI is co-directed by S. Shankar Sastry, a professor of electrical engineering and computer science at the University of California, Berkeley, and Rayadurgam Srikant, a professor of electrical and computer engineering at the University of Illinois, Urbana-Champaign. Both of them have extensive experience with the supercomputers at their respective institutions, according to The New York Times.

“The probability of something good not coming out of this is zero,” Thomas Siebel, CEO and co-founder of C3.ai, explains to the Times.

Applications for access to the supercomputer network are now being accepted. C3.ai DTI anticipates providing 26 grants worth up to $500,000 each year. Results of projects that make use of the computers are required to provide public access to the results. To apply for a grant, the principal investigator of the project must be at one of the partnering universities, but team members can be brought aboard from any location.

“Ultimately, we need a cure. To be able to tackle that, we need to accelerate science,” Dario Gil, director of IBM Research, tells the WSJ. “We do have an extreme sense of urgency.”

Read full article here.