Please feel free to forward this invitation to anyone you think might be interested.
Feb. 6, 2025 at 3PM (EST)
Adam Knapp, University of Florida. Data Assimilation for ABMs using the Ensemble Kalman Filter
Constructing a Medical Digital Twin will require the use of data assimilation (DA) techniques, in which measurements from the patient are incorporated into the computational model. However, many computational models of disease are based on ABMs, which violate a number of the assumptions underlying existing DA techniques. I will present a technique to get around these difficulties by working on two levels: the micro- and macro-state. This distinction allows us to implement a DA technique commonly used in weather forecasting, the Ensemble Kalman Filter, on the vector-valued macrostate level, which is coupled to a corresponding update at the microstate level.