Title: Data-driven deep learning algorithms for time-varying infection rates of COVID-19 and mitigation measures

Abstract: Epidemiology models with constant parameters may not capture the infection patterns in the presence of pharmaceutical and non-pharmaceutical mitigation measures during a pandemic, since infectiousness is a function of time. In this paper, Epidemiology Informed Neural Network (EINN) algorithms are introduced to discover time-varying infection rates for the COVID-19 pandemic. Since there are asymptomatic infectives, mostly unreported, EINN learns the probability that an infective individual is asymptomatic. Using cumulative and daily reported cases of infectives, we simulate the impact of non-pharmaceutical mitigation measures such as early detection of infectives, contact tracing, and social distancing on the basic reproduction number. We demonstrate the effectiveness of vaccination, a pharmaceutical mitigation measure, together with non-pharmaceutical mitigation measures on the daily reported infectives. The EINN algorithms discover time-varying infection and recovery rates. The Mean Squared Error is used to demonstrate the accuracy of the proposed EINN algorithms. Simulations are presented for Italy, South Korea, United Kingdom, and the United States.