Their working paper (ziadobermeyer.com/wp-content/uploads/2020/07/…) combines group testing and machine learning to make repeatedly testing entire populations for COVID feasible. It's a nice blend of medicine, data, and tech.
@M.A.R. The basic idea is to mix samples from a group smartly selected by machine learning so that one test can rule out a large number of people. If it's positive, everyone in the group gets retested and you quarantine the infected, making it more likely to get negative tests...