[Colloq] Reminder : Guest Lecture : Computational Bayesian Estimation with Applications in Sensors and Tomography; Fri April 27 10 - 11am 366WVH

[Ponte, Christopher]

Speaker:              Clemens Heitzinger
Date:                    Friday,  April 27, 2018
Time:                   10:00-11:00 am
Location:             366 WVH
Title:                    Computational Bayesian Estimation with Applications in Sensors and Tomography

Abstract:
We use Bayesian estimation to extract as much information as possible in two sensing applications, namely in nanoscale field-effect sensors and in electrical-impedance tomography.  Computational Bayesian estimation provides us with the means not only to estimate unknown parameter values, but also their probability distributions and hence uncertainties in reconstructions.  This approach also includes a physical model; in both applications considered here, the physical models are partial differential equations.

We present our algorithm for multi-dimensional Bayesian estimation and discuss numerical results for the (ill-posed) inverse problem in electrical-impedance tomography, showing that certain properties of the internal objects are much harder to reconstruct.  In the application of a nanoelectrode field-effect sensor, we characterize the devices and show how various measurements at different frequencies can be combined to increase accuracy.

Whereas previous approaches relied on classical methods in parameter estimation or inverse modeling and could not solve the reconstruction problem for electrical-impedance tomography, our new algorithmic approach is a much more principled way to incorporate information and also allows to decide whether a reconstruction problem is solvable or not. At the end, we shortly discuss how the general idea can be applied to far broader settings where model parameters need to be determined indirectly, relevant to many problems in science and engineering.
About the Speaker:
Clemens Heitzinger received his master's and PhD degrees from TU Wien. He was a visiting researcher in the Department of Mathematics and Statistics at Arizona State University, a research associate in the School of Electrical and Computer Engineering at Purdue University, and a senior research associate in the Department of Applied Mathematics and Theoretical Physics (DAMTP) at the University of Cambridge.  In 2015, he returned to TU Wien as an Associate Professor in the Department of Mathematics and Geoinformation.  He was awarded the START Prize by the Austrian Science Fund (FWF), Austria's most prestigious award for young scientists.  His research interests are uncertainty quantification (stochastic partial differential equations and Bayesian estimation) and machine learning.

http://Clemens.Heitzinger.name/<http://clemens.heitzinger.name/>