Michael Chertkov
University of Arizona
Library to solve Inference, Learning and Optimization problems (related to COVID-19 and more general) based on Graphical Model approach.
This project will focus on design and testing of Reduced Models. Will need "ground truth" real and synthetic data. On the latter - very much interested to get synthetic data from from agent- and individual- based simulations on the scale of a campus/county/city/state (which will be a verifiable "ground truth" test for us)
Advantage of our approach is in ability to integrate diverse data and constraints into one comprehensive model providing probabilistic prediction on how COVID-19 spreads. The data and constraints may be linked to epidemiology, social behavior, geography, etc. We are interested to get input on different sources of the data and constraints to be accounted for in our integrated model from other RAPID program contributors.
reduced models graphical models machine learning inference optimization