By Chris Hayhurst, Contributor
Few educational theaters can compare with the Frost Planetarium at the Phillip and Patricia Frost Museum of Science (Frost Science) in Miami, Florida. Set by the shore near downtown, the 250-seat facility, with its 67-foot dome ceiling, deploys surround sound, laser lighting, and a 16-million-color 8K projection system to create an immersive experience that draws visitors from across the country.
But late this March, when throngs of children normally would have been waiting for the “Journey to the Stars” show to begin, this technological marvel wasn’t open for business. Instead, it had been devoted to an even higher purpose: helping in the fight against the coronavirus (COVID-19).
As the outbreak gained steam (Florida residents were ordered to stay home), Frost Science shut its doors “just like everyone else,” explains Brooks Weisblat, the museum’s vice president of technology. “But then we started thinking about what we could do to help, and how all this technology we had was essentially idle, and we realized we could probably put it to use.”
They soon discovered the Berkeley Open Infrastructure for Network Computing (BOINC), a crowd-sourced computing platform funded by the National Science Foundation and implemented by the University of California, Berkeley. The platform pools unused computing resources from all over the world, and then makes them available to individual researchers focusing on specific projects. One such initiative, “Rosetta@home,” led by a lab at the University of Washington’s Institute for Protein Design, is using the platform to study coronavirus proteins.
“Basically, they’re trying to answer these really complex questions to develop antiviral drugs and possibly a vaccine, and they need all the computing power they can get.”
—Brooks Weisblat, vice president of technology, Frost Science
The museum is making Frost Planetarium’s Dell EMC PowerEdge servers available for the Rosetta@home project, which consists of processors, Dell EMC SC Series storage and Precision with Nvidia Quadro graphics cards.
“Basically, they’re trying to answer these really complex questions to develop antiviral drugs and possibly a vaccine, and they need all the computing power they can get,” Weisblat says. “So now our servers are all cranking away for them instead of just sitting here with nothing to do.”
Tech Attack: Diagnosing, Tracking, and Treating Coronavirus
While social distancing, thorough hand washing, and the heroic efforts of medical professionals are all key to ending the coronavirus pandemic, many believe technology has a role to play, as well—and may ultimately drive long-term recovery and success.
Among the companies and institutions doubling down on technologies they hope can help stem the crisis: the Massachusetts Institute of Technology, which has a group developing a design for a DIY ventilator that hospitals should be able to build with $100 worth of parts; a firm in Singapore called MaNaDr that’s built a telehealth app already used by more than a million patients; and the “HealthMap” team at Boston Children’s Hospital, which is using machine learning to scan social media posts and other publicly available sources of data and information to map the spread of coronavirus in real time.
Elsewhere, an Israel-based start-up called Sonovia is developing a method for embedding fabrics with antiviral nanoparticles that may prevent coronavirus transmission. Steth IO, of Bothell, Washington, has created a listening device that can be used with an app to detect COVID-induced buildup of fluid in the lungs.
Artificial Intelligence Leads the Way
If one technology has been deployed more than others in the fight against COVID-19, it would have to be artificial intelligence (AI). Some hospitals, for example, are using AI algorithms integrated into their electronic health record (EHR) systems to identify patients with coronavirus who are at high risk of deterioration, while others have used the technology to build online, chat-bot screening tools that differentiate between probable COVID-19 patients and those with other health concerns. The Alphabet subsidiary DeepMind, meanwhile, has a system it’s dubbed “AlphaFold” that uses AI on genomic datasets to generate 3D models of protein structures. And the Maryland-based Insilico Medicine is leveraging AI to identify molecules that might be used to create coronavirus therapeutics.
Typically, novel drug discovery takes three to five years, notes Nathan Collins, Ph.D., chief strategy officer in the biosciences division of the nonprofit research institute SRI International. “Well, that’s clearly too long to wait when you’re talking about something like COVID-19,” he says. SRI’s goal: Working in collaboration with a company called Iktos, which specializes in using AI to design drug compounds quickly and efficiently, they hope to soon have an automated system that streamlines the process dramatically. “We’re aiming to reduce it to about six months,” Collins says. “And that could get even faster as the technology develops.”
At SparkBeyond, data scientists from around the globe decided to use its technology to build geographical “heat maps” that show where infected patients have been, identify asymptomatic carriers, and predict future high-risk areas.
An AI-focused organization with similarly lofty ambitions is SparkBeyond, a New York City-based company brandishing the tagline, “Unleash AI for Impact.” The firm has always been interested in “social-good-type projects,” says Ryan Grosso, director of data science. Once it became obvious that COVID-19 was serious, SparkBeyond jumped at the opportunity to make a difference. “We have this extremely powerful platform that lets us develop actionable insights from complex sources of data,” he explains. “So we all got together”—they have 50 data scientists globally—“and we started brainstorming different ideas.”
In the end, the team decided it would use its technology to build geographical “heat maps” that show where infected COVID-19 patients have been, identify asymptomatic carriers, and predict future high-risk areas. The big challenge moving forward, Grosso explains, is convincing government agencies, including those in the U.S., to be more forthcoming with their data.
In regions of the world where data are widely available—Grosso points to Israel and Italy as examples—SparkBeyond’s data scientists have leveraged their platform to great effect. For one recent project, the company’s algorithms identified public drinking fountains, gas stations, and other areas in Milan correlating with higher rates of infection. In another, for Tel Aviv, they created a heat map for city officials providing “building-level resolution” of the virus as it spread.
SparkBeyond’s AI platform, like many other emerging technologies, is designed to help experts like public health officials make “very fast and informed decisions,” Grosso says. In a situation as serious as the current one, he adds, having such tech available is not only useful, “it’s exactly what you need if you’re trying to save lives.”