The systems we rely on to forecast and track the spread of infectious diseases could vastly improve, thanks in part to work by a University of Virginia AI researcher.
Chen Chen , an assistant professor of computer science at the University of Virginia’s School of Engineering and Applied Science, is developing trustworthy, next-generation AI systems that address challenging real-world problems in fields like public health. Her current project focuses on a new approach to epidemic intelligence, building computational tools that combine multiple sources of information — from reported cases and wastewater surveillance to human travel data — and using artificial intelligence to better predict where diseases will spread and help public health officials know how to respond.
The work grew from Chen’s Ph.D. research on connectivity in complex networks. Her dissertation developed a unified framework for measuring and understanding network connectivity across different application systems, ranging from infrastructure networks, such as power grids, to epidemic networks. In these networks, she said, the behavior of an entire system depends on how its individual nodes are connected.
“This research focus is a natural extension of that idea,” Chen said. “People are all connected — we are connected to our co-workers and families.”
If a node fails in a power grid, engineers want to understand whether other parts of the network can keep the system operating. In an epidemic, the nodes are people, and their connections can become pathways for disease transmission across a specific geographic area. In a pandemic, the transmission of that disease skyrockets, expanding exponentially and crossing international boundaries.
“For the epidemic part, we are focusing on how to effectively contain the spread of disease. We want to come up with a strategy to monitor the whole population outbreak at different levels,” Chen said. “Ultimately, we want to contain disease transmission and reduce the number of infections.”
That research has earned Chen a five-year, $600,000 CAREER Award — the National Science Foundation's Faculty Early Career Development Program award — for her project, “ Building Next-Generation Epidemic Intelligence: Forecasting, Intervention, and Surveillance .” The CAREER program supports early-career faculty who have the potential to serve as academic role models in both research and education.
Existing epidemic forecasting models and tracking models can miss important pieces of the picture, Chen said. A model might capture reported infections, for example, without fully incorporating how people are moving between communities or what wastewater monitoring suggests about infections that have not been reported.
“The mobility of people matters a lot,” she said. “Because if you are traveling a lot from state to state, you would establish connections with different people — and this may pose infection risk as well.”
The emergence of analytical tools powered by artificial intelligence is only now allowing the kind of data processing these projects require, said Chen, a former Google software engineer, whose industry experience influenced how she approaches these problems. It’s not only about developing new algorithms, she said, but thinking about how they can scale to large, complex datasets and ultimately be useful in real-world settings.
Chen wants to bring those data sources together in real time, and her research interests in data mining, machine learning, computational epidemiology, and trustworthy and efficient AI will help her do that in a way that aids both scientific discovery and public health decision-making.
Trustworthy AI, Chen said, is particularly important in high-stakes applications such as public health, where AI models need to account for uncertainty and potential errors and provide results that researchers and decision-makers can understand and rely on.
Disease itself can also change. The emergence of new variants during COVID-19 demonstrated how quickly the characteristics governing the disease and how it is transmitted can shift. Chen’s goal is to develop models that can adjust as such conditions change, said Guanghui Min, a Ph.D student working with her on the project.
“One exciting next step is to combine different signals, such as case counts, mobility patterns and wastewater data, to get a more complete picture of how an outbreak is evolving,” Min said. “We think making these models adapt when conditions change so that forecasts can help identify where additional monitoring or intervention is needed may be most useful.”
The underlying approach to Chen’s research would not be limited to COVID-19. By changing how a model represents transmission and contact patterns, Chen said, the same framework could potentially be adapted to different infectious diseases.
The CAREER project has several components: improving epidemic forecasting, developing computational tools to help evaluate possible interventions, designing better disease-surveillance strategies, and educating the public. Chen plans to concentrate heavily on forecasting during the first three years, creating a foundation for the intervention and surveillance tools that follow.
Those tools could help public health officials answer practical questions: What is happening now? What is likely to happen next? And if officials intervene, how might the trajectory change?
“We live in interesting times in that we have access to unprecedented amounts of data, analytical tools, and real-world insight from COVID-19 about how to apply them,” said Sandhya Dwarkadas , Walter N. Munster Professor and chair of the Department of Computer Science. “I appreciate how Chen is maximizing this advantage to help make a difference for communities worldwide by informing public health decision makers.”
For Chen, the ultimate measure of success is whether better computational models can turn scattered information into useful guidance — helping officials decide how to respond to epidemics, prevent pandemics and help individual community members better understand the risks around them.
A central challenge is scale. Inside a hospital, researchers might be able to map interactions between infected individuals in detail. Across a city, state or country, that becomes far more difficult.
Chen’s approach is designed to work with different levels of available information. Publicly available disease reports, wastewater measurements and mobility data could provide a broader view of disease spread. When more detailed information — such as localized case counts, testing data and other health surveillance records — is available for research, the framework could incorporate it to provide a more granular understanding of transmission and improve predictions.
The research could eventually help identify where additional monitoring would have the greatest impact. If a region lacks sufficient information, for example, a model might show public health officials where another wastewater monitoring point could most improve forecasts.
In addition to computational models and decision-making tools for public health officials, Chen also envisions building educational tools for the public. Her long-term goal is an accessible website that could help members of the community understand disease conditions and risk in the places around them.
“Our ultimate goal is to do something like that — so people can know the overall condition of nearby areas,” Chen said, referring to public-facing disease-tracking websites. “What is the viral level there? What is a risk level if you are visiting that place?”
“Ultimately, we want to come up with a strategy to monitor the whole population outbreak at different levels,” Chen said.