Seminar series
|
|
|
|
Session 2005-2006 Spring Summer Session 2006-2007 Autumn Spring Summer Session 2007-2008 Autumn Spring Summer Session 2008-2009 Autumn Spring Summer Session 2009-2010 Autumn Spring Summer Session 2010-2011 Autumn Spring Summer Session 2011-2012 Autumn Spring Summer
|
Leverhulme LectureComputational Methods for Quantifying Uncertainty in Biological ModelingProfessor R. Mike Kirby (University of Utah)1400, Tuesday 17th February, 2009Room 9.30a, School of Computing, University of LeedsAbstractNumerical simulations entail modeling and approximation assumptions that impact outcomes. Choices concerning items like geometric models and parameters made as part of the numerical simulation process may significantly influence the accuracy of the results, and hence evaluating the sensitivity of simulation results to variations in these choices is essential. Although the field of sensitivity analysis is well developed, systematic application of such methods to complex biological models is limited due to the associated high computational costs and the substantial technical implementation challenges. In this talk, we present three case studies in which the use of the generalized Polynomial Chaos (gPC) technique offers a computationally feasible alternative to traditional Monte Carlo approaches for assessing the impact of model and parameter variability. In the first study, we examine the impact of torso conductivity parameter values on the forward problem of electrocardiography (ECG). In the second study, we examine the impact of heart position on the forward ECG problem. In the third study, we propose a framework that combines component shape parameterizations with the stochastic approximation methods to study the effect of drug depot shape variability on the outcome of drug diffusion simulations in a porcine model. In all three cases, we demonstrate that once the underlying stochastic process is characterized, quantification of the introduced variability is quite straightforward and provides an important step in the validation and verification process.The SpeakerRobert M. (Mike) Kirby is an Associate Professor of Computer Science, Adjunct Associate Professor of Mathematics and Adjunct Associate Professor of Bioengineering at the University of Utah. He is both a faculty member in the School of Computing and the Scientific Computing and Imaging Institute at Utah. He received the Sc.M. degree in Applied Mathematics, the Sc.M. degree in Computer Science, and the Ph.D. degree in Applied Mathematics from Brown University, Providence, RI, in 1999, 2001, and 2002, respectively. His research focus is on large-scale scientific computing and visualization, with an emphasis on the scientific cycle of mathematical modeling, computation, visualization, evaluation, and understanding. His primary research interests include: Computational Science and Engineering Applications, Algorithm Development and Application of High-Order Methods, Scientific Visualization, Verification and Application of Concurrent Programming, and High Performance Computing. Leverhulme Visiting ProfessorshipThe objective of the awards is to enable distinguished academics based overseas to spend between three and ten months at a UK university, primarily in order to enhance the skills of academic staff or the student body within the host institution. It is recognised that Visiting Professors may also wish to use the opportunity to further their own academic interests. The over-riding criteria for selection are first the academic standing and achievements of the visitor in terms of research and teaching, and second the ability of the receiving institution to benefit from the imported skills and expertise. Priority will be given to new or recent collaborative ventures.
|