Events Calendar

ISE 651 Epstein Institute Seminar

Speaker: Dr. Matthew Plumlee, Assistant Professor - University of Michigan

Talk Title: Bayesian calibration of complex systems: methods, applications, and lingering issues

Abstract: -“ Bayesian calibration is used to combine collected
experimental or observational data with existing
understanding. The popularity of this method is tied to several
features: it can incorporate unknown parameters, it can function
when the model is imperfect, and it offers both future prediction
and uncertainty quantification. This seminar will discuss the
general methods and some applications of Bayesian calibration
while challenging the current state-of-the-art science on the
topic. Chiefly, this talk will discuss a modeling framework that is
widely adopted but leaves the posterior of the parameter often
sub-optimally broad. There has been no generally accepted
alternatives to date. This presentation will illustrate how using
Bayesian calibration where the prior distribution on the bias
(potential model incorrectness) is orthogonal to the gradient of the
computer model. Problems associated with Bayesian calibration
are shown to be mitigated through analytic results in addition to example.

Biography: Matthew Plumlee received his Ph.D. degree in Industrial and Systems Engineering from the
Georgia Institute of Technology. He is currently an Assistant Professor in the department Industrial and
Operations Engineering at the University of Michigan. His research interests include statistical learning with
special emphasis on model calibration and uncertainty quantification. He received the Georgia Tech Sigma Xi
Best Dissertation Award and the Quality, Reliability and Statisitistic Best Student Paper Award from INFORMS.
His work has appeared in the Journal of the American Statistical Association and Technometrics.

Host: Dr. Jong-Shi Pang & Dr. Qiang Huang

Location: Ethel Percy Andrus Gerontology Center (GER) - 206

Audiences: Everyone Is Invited

Contact: Angela Reneau