You are here

Statistics Seminar: Oksana Chkrebtii

Applying Principles of Statistical Design to Adaptive Probabilistic Differential Equation Solvers
November 7, 2019 - 11:00am

Lind 305

Seminar Speaker: Oksana Chkrebtii, Department of Statistics, The Ohio State University

Title: Applying Principles of Statistical Design to Adaptive Probabilistic Differential Equation Solvers

Abstract: When models are defined implicitly by systems of differential equations with no closed-form solution, small local errors in finite-dimensional solution approximations can propagate into deviations from the true underlying model. Some recent perspectives in quantifying this uncertainty are based on Bayesian probability modeling: a prior is defined over the unknown solution and updated by conditioning on interrogations of the forward model. We apply principles of Bayesian statistical design to sequentially adapt time-steps for state-space probabilistic differential equation solvers and investigate the behaviour of local error under the adaptive scheme which underlies numerical variable step-size methods. 

Bio: Oksana Chkrebtii joined the Department of Statistics at The Ohio State University in 2014. Her research focuses on inference for models of the spatio-temporal evolution of states that incorporate known physical and biological laws. 

Oksana Photo