May 23, 2022
Linda Blaser and Meghan O’Toole
Professors collaborate to model the COVID-19 pandemic
Using publicly available data, Assistant Professor of Chemistry Will Conrad and Assistant Professor of Mathematics Arthur Bousquet combined traditional epidemiological approaches with machine learning to better predict the course of the COVID-19 pandemic in France, Germany, the UK, and South Korea.
The new approach improved the ability of researchers to predict the direction of the pandemic within seven days. Their research was published in the journal, Scientific Reports, under the title “ Deep learning forecasting using time-varying parameters of the SIRD model for Covid-19” and included research contributions from students Said Omer Sadat ’21 and Nelli Vardanyan ’22 as well as their South Korean collaborator, Assistant Professor of Mathematics Youngjoon Hong from Sungkyunkwan University.
“Pandemic modeling is very important for both policymakers and stakeholders to decide what actions are safe. Should we impose a lockdown? Should we wear masks in public? Should we re-open? Our model increases the accuracy of existing models for more refined decision-making. This saves lives and money,” Conrad said of the study’s impact.
Bousquet and Conrad began the study in the summer of 2020 and have collaborated on projects since joining the faculty in 2018. They value the liberal arts and “the power of interdisciplinary collaboration that a place like Lake Forest College brings,” they said.
Students Sadat and Vardanyan were enlisted to help with the research. The interdisciplinary nature of the study is what appealed to them.
“Said was interested in data science and was also trying to get into medical school, so he wanted a project where data science could apply to biology,” Bousquet said.
Analyzing pandemic data to predict infection rates in a set population presented an ideal opportunity for such interdisciplinary work that was also presently relevant.
“We saw it as an ‘all hands on deck’ moment to fight the coronavirus pandemic,” Conrad said. “We wanted to use whatever skills we had to be able to contribute. We developed this project with that idea in mind.”
The study also demonstrated new possibilities for the use of machine learning in the analysis of data.
“In the field of data science, we solve partial differential equations in different ways, and in this study, we incorporated machine learning, which is what was new for us,” Bousquet said of the study’s greater impact. “We demonstrated how to use machine learning to solve a partial differential equation.”
Collaboration between faculty members generates unique and interesting opportunities for interdisciplinary research. Conrad and Bousquet have developed a collaborative relationship that has benefitted learning by exposing students to the intersections of different fields.