Since the COVID-19 High Performance Computing Consortium was formed at the onset of the pandemic to open up access to supercomputers for researchers across the planet, it advanced more than 90 coronavirus-combatting research projects and almost doubled its available computing capacity.
Now, the group is moving on to the next chapter.
The public-private partnership, launched in March by the White House, Energy Department and IBM, entered what insiders on Monday deemed its “second phase,” which will prioritize research projects that hold promise to help boost patient outcomes in the next half-year.
“After eight months, we’ve certainly learned a lot about what we can do with the consortium and the kind of research that it’s been able to generate,” IBM Systems General Manager, Strategy and Development Jamie Thomas told Nextgov Monday. She added that this latest move marks a sense of “renewed focus” aimed at driving additional proposals and interest.
Thomas has had a long tenure at IBM, where she first started as a computer programmer years ago. She’s now managing a large innovation team inside of the company that brings products to the market, including a lot of the hardware underpinning some of the supercomputers the consortium uses. With a front seat view to the ongoing work, she offered a glimpse into how it’s evolved so far, potential research the second wave will incorporate, and hopes that the broader effort might pave the way for a national “science reserve” in the future.
From Phase I to II
The COVID-19 HPC Consortium came together swiftly in March, not long after the first wave of COVID-19 rattled health care systems, supply chains, and roughly every aspect of American life. At the time, a relatively smaller group of participating national laboratories, agencies, companies and academic institutions volunteered to share free compute time and resources—encompassing some of the world’s most powerful systems—for outsiders to use, all to help put an end to the global health emergency.
Participation continues to grow and the consortium is now made up of 43 members spanning the U.S. public and private sectors, academic institutions, as well as several international government agencies. As more and more partners hop on board, the resources available to researchers and the overall computing capacity are notably expanding.
The consortium launched with an initial computing capacity of 330 petaflops in March, which it has since almost doubled to 600 petaflops. The hope has been that such power might help scientists—who can apply online any time to leverage the machines—deliver new research results in hours or days, instead of weeks or months.
“There’s now 175 different research projects that have entered into the consortium, so that’s a pretty interesting learning point,” Thomas explained. “We also do have a lot more patient data—anonymized data, of course—available to now utilize in the context of what we’re looking at.”
The chosen, already in-the-works and active 91 projects harness the massive computing capabilities for a wide range of investigations. Early on, they included studies that introduced artificial intelligence and machine learning to advance new discoveries around drugs and therapeutics that can be used against the virus, one that uses a NASA-run supercomputer to define risk groups for severe pulmonary disease associated with COVID-19, another that supported the design of a multi-splitting device for ventilators to help coronavirus patients—and many more.
“In the beginning, it was kind of an open landscape,” Thomas said.
But in these next few months, there will be a sharper focus on research that can drive speedy, patient-centered treatments and outcomes.
“The second phase of the COVID-19 High Performance Computing Consortium can potentially provide tangible results to those affected by the virus, and we look forward to delivering these results to the American people,” Energy’s Undersecretary for Science Paul Dabbar said in a statement Monday.
Research proposals will continue to be accepted on the consortium’s online portal. Officials will specifically—but not exclusively—hone in on proposed projects that involve: modeling and grasping patients’ responses to the virus using large clinical datasets; verifying vaccine response models from various clinical trials; assessing therapies with repurposed molecules; and “epidemiological models driven by large multi-modal datasets,” the release said.
Thomas also confirmed that patient “privacy is very important” to those involved in the effort and that the data used is anonymized. It comes from a range of sources, depending on the reach of institutions. “A lot of these universities have different relationships with different organizations that they’re able to get the data from,” she noted.
This boost in data volumes since the consortium’s launch is part of what’s influenced the newly announced transition.
To help visualize how the focal point is shifting toward driving patient treatments and quicker outcomes, Thomas pointed back to the announcement, which highlights research that she said represents what they’re “looking to take forward” in the next phase. Offering an example, she said a research team from Utah State University used consortium computing power to simulate the dynamics of aerosols indoors, which shed new light on how droplets from humans’ breaths may linger longer in the air. Separately, but significantly, a research group from Iowa State University conducted research on certain genes that could help inform why some minorities are more vulnerable to, and adversely affected by COVID-19.
“Understanding the effects on different populations is very important—and understanding whether that’s genetic or environmental, or other aspects, is key I think to containing and mitigating for those for those individuals,” Thomas said.
What Could Come from the Consortium
Considered an entry mechanism for groups in academia, agencies or industry that might have a need for large-scale computing capabilities, the COVID-19 HPC Consortium offers, for many, a new, simpler way to access it.
Thomas noted that that wasn’t always the case.
“Before this, the reach into some of these resources was not as prescribed as it is now through the consortium—[where] organizations are able to present their research and have a board process it, then determine whether it warrants a use of this capacity,” Thomas said. “So I think it’s really an important change in terms of how these researchers are able to have democracy, if you will, democratic access to these resources.”
America is inching closer in the development of usable vaccines, but patient-focused treatments are also “vital,” Thomas noted. Once the next phase is completed, there might be more. Or, she said, “hopefully COVID-19 is going to be a closed case—but there’s many other things that we could use these resources to solve.”
Looking ahead to how this all might be helpful in the future, Thomas pointed to an open letter IBM’s CEO recently penned to the incoming Biden administration, to, as Thomas put it “basically say, ‘science is really important to the country and to the world, and maybe we should consider the notion of a science reserve.’”
How such a reserve would be formulated is “to-be-defined,” she said, but at the highest level, Thomas envisions it as a hefty investment made by the government to allow organizations to more easily construct these kinds of broader consortia to collectively tackle big problems beyond the pandemic—like climate change.
“Also, we very much feel that technology should be used to serve humans in a responsible manner, and this is really a responsible use of technology to help individuals,” Thomas said. “That, I think, is a key takeaway from this whole effort, and something that we need to continue going forward.”