COVID-19 Forecasting

I was a research intern in the COVID modeling team at Wadhwani AI and a volunteer with the COVID-19 Data Science Consortium. Our team performed medical demand forecasting and scenario modeling for highly impacted Indian regions. We developed SEIR-like compartmental models for forecasting, using Bayesian optimization for parameter estimation (source code). Our work was accepted at ACM CODS-COMAD 2021 and at the ICLR 2021 Workshop on AI for Public Health. We also submitted forecasts to CDC through the Reichlab COVID-19 Forecast Hub (paper).

Along with researchers from Data Science India vs COVID-19, I also worked on the problem of managing an epidemic by determining the optimal policy for transmission control. To this end, we proposed an analytical framework combining aspects of epidemiological models (SIR) and Lotka-Volterra (LV) systems (pre-print) (source code). Our work was accepted at the ICLR 2021 Workshop on Machine Learning for Preventing and Combating Pandemics.