Leadership in providing with the public consistent projections of COVID-19 infections and hospitalizations has prompted the launch of a USF-led global modeling project sponsored by Microsoft that can ultimately help scientists more swiftly identify and respond to the threat of emerging infectious diseases.
Edwin Michael, professor of epidemiology in the USF College of Public Health, has been tapped to build a smart learning-based computational system that creates a library of simulation models that are updated in real-time as it pools data provided by public health sources and journal publications from around the world. In collaboration with USF Information Technology, Michael was awarded a grant from Microsoft's AI for Health program to utilize Microsoft Azure to develop a software system that teaches itself how to recognize the individual characteristics of airborne, waterborne and mosquito-borne diseases. It will take socioeconomic and ecological structures into account, which vary by location and could impact viral transmission of a particular disease. For example, cholera is typically contracted from contaminated water in impoverished regions with poor sanitation.
鈥淭he COVID-19 pandemic has highlighted how understanding and predicting emerging and other outbreak diseases belong to modeling, decision and policy domains that are marked by significant uncertainty, complex global interdependencies and heterogeneity,鈥� Michael said. 鈥淲e are excited about the prospect of combining systems dynamics modeling with Microsoft鈥檚 data science, large-scale cyberinfrastructure, artificial intelligence and visualization strengths for developing a novel, agile, 鈥榮mart鈥� process-based computational system to address this challenge, starting with the Microsoft Azure Services.鈥�
Artificial intelligence will be utilized to establish a system of mathematical models that track the lifecycle of a particular virus 鈥� infection, symptoms, hospitalization, intensive care and death, helping guide the decision-making process for policymakers, public health officials, health care providers, businesses and schools. The data will play a key role in assessing a virus鈥檚 transmissibility, prediction of its future course and the potential for mitigation, containment and eventually control of a pandemic.
鈥淲ith COVID-19, we have seen firsthand the role that cloud computing can play in helping advance research quickly, securely and efficiently,鈥� said Jamie Harper, vice president of U.S. Education for Microsoft. 鈥淲e support the work at the 91社区 and look forward to the advances and learnings that will continue to spur innovation.鈥�
The modeling architecture will be accessible via Github, in which all code, data, services and publications are open source. This allows new and evolving information regarding the transmission and control of a virus to be rapidly incorporated into the models, helping quantify the likely impacts of new virus strains and emerging interventions, such as vaccinations and therapeutics. Through its development, Michael will build cross-campus collaborations between disease modelers, epidemiologists, data scientists, engineers, mathematicians and social scientists and launch interdisciplinary graduate and post-graduate student training programs. He expects to complete the first prototype in one year.
"IT brings industry partners and researchers together to leverage modern cloud technologies including serverless architecture and AI to allow for rapid innovation,鈥� Sidney Fernandes, vice president of Information Technology and chief information officer, said. 鈥淭his strategy allows USF researchers to take advantage of the flexibility and speed of the cloud and is a game changer for researchers like Dr. Michael. We were thrilled to partner with Microsoft and Dr. Michael to enable this critical project."
The intelligent disease modeling tool builds upon Michael鈥檚 recent work in developing the SEIRCAST simulation portal, which county and state health officials have utilized since mid-2020 to plan for COVID-19 outbreaks in Tampa Bay and a surge in hospitalizations. His expertise also helped guide decisions made at USF regarding campus operations. Michael is now working on a new projection model for Hillsborough County that looks at vaccination rates by zip code, helping pinpoint where hotspots may arise. He plans to have his first set of data available within the next few weeks.