Ahmed M. Alaa

Appendix for "When and How to Lift the Lockdown? In this Section, we provide a detailed comparison between different existing approaches for modeling fatality curves. A tabulated comparison between our model and existing ones is laid out in Table A1. Table A1: Comparison between existing approaches to model fatality curves. Because of the relative infrequency of pandemics, little related work has been done within the machine learning community to address this problem. In what follows, we provide a brief overview of previous works. Previous works prior to the current pandemic have been primarily focused on learning contagion (diffusion) processes on networks, e.g., [15, 16]; unfortunately, these models do In response to the COVID-19 pandemic, two strands of research work have emerged: (a) methods for devising optimal control (lockdown) policies to contain disease spread [17, 18, 19], and (b) models for forecasting disease spread and expected fatalities [5, 6, 7, 9, 10, 20, 21].