AI Harnessed in COVID-19 Battle

Tech data could help narrow the pandemic knowledge gap.

As COVID-19 grips the globe, government and institutional leaders are making major decisions with potentially serious consequences. Closing schools, mandating remote work, and crafting legislation to salvage crippled economies indicate a balancing act between saving lives and preserving livelihoods. A major part of the difficulty is the dearth of information about this novel coronavirus, SARS-CoV-2—its qualities and behavior, and how to stop it. From the early weeks of the outbreak, researchers have turned to artificial intelligence (AI) to help diagnose the disease, track its spread, and find treatments for it.

Bluedot’s Global Early Warning System

Possibly the first entity to sound the alarm on the outbreak of the respiratory disease in Wuhan, China, was Bluedot, an infectious disease surveillance company that uses an AI algorithm to track information released around the world for signals of disease spread. The company alerted its clients before the Centers for Disease Control and Prevention or the World Health Organization announced the emergence of the virus. Bluedot’s algorithm relies on machine learning to pick up on possible outbreaks from a host of information sources. Epidemiologists then review the AI results to determine a course of action. Another company, Metabiota, was also an early detector of COVID-19.

A Chinese Supercomputer Offers an Alternative Diagnosis Strategy

Housed at the National Supercomputer Centre in Tianjin, a powerful Chinese supercomputer offers an alternative to the commonly used nasal swab test for COVID-19. The Tianhe-1 computer runs an AI diagnostic tool, which analyzes CT scans of patients for signs of the disease. So far, health researchers have found that this coronavirus leaves tell-tale marks on patients’ lungs—a sign that could help ramp up testing given the lack of conventional test kits in many countries. Tianhe-1 reportedly can analyze hundreds of scans and give diagnoses within 10 seconds, with accuracy above 80 percent.

IBM Supercomputer Searches for Cures

Known as “the world’s most powerful supercomputer,” IBM’s Summit supercomputer at the U.S. Department of Energy’s Oak Ridge National Laboratory has been sifting through and analyzing drug compounds, trying to find ones that can be used against SARS-CoV-2. Summit possesses 9,216 IBM Power9 CPUs that are able to crunch 200 quadrillion computations per second. This ability has so far been used to study 77 drugs in two days, searching for ones that will render the virus unable to infect host cells.

Google’s Deepmind Illuminates SARS-CoV-2

Google’s AI system, Deepmind, has been employed to gain a better understanding of the novel coronavirus itself. Part of the AI’s work is to predict how the viral proteins develop using its AlphaFold system. This knowledge could help researchers understand and predict how the virus will behave in the future, such as how it may spread or mutate.

BenevolentAI Analyzes Existing Drug

By using a repertoire of AI algorithms, BenevolentAI has identified Baricitinib as a possible drug that could be effective against SARS-CoV-2. BenevolentAI identified a group of drugs that could stop the viral infection of cells based on in vitro evidence, and concluded that Baricitinib is the best candidate. The company has suggested that clinical trials move ahead for this drug to potentially counter COVID-19. The drug is currently prescribed to treat rheumatoid arthritis.