BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Reduced order modeling for uncertainty quantification in cardiac e
 lectrophysiology - Andrea Manzoni (Politecnico di Milano)
DTSTART:20190606T093000Z
DTEND:20190606T100000Z
UID:TALK125650@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:We   present a new\, computationally efficient framework to pe
 rform both forward   and inverse uncertainty quantification (UQ) in cardia
 c electrophysiology. We   consider the monodomain model to describe the el
 ectrical activity in a   subject-specific left ventricle geometry\, couple
 d with the Aliev-Panfilov   model to characterize the ionic activity throu
 gh the cell membrane. We take   into account relevant inputs related to bo
 th models\, such as electrical   conductivities\, pacing times\, and coeff
 icients affecting the ionic models. We   address a complete UQ pipeline\, 
 including: (i) a variance-based sensitivity   analysis for the selection o
 f the most relevant input parameters\; (ii)   forward UQ to investigate th
 e impact of intra-subject variability on   clinically relevant outputs rel
 ated to the cardiac action potential\, and   (iii) inverse UQ for the sake
  of parameter and state estimation within a   Bayesian framework. All thes
 e stages exploit stochastic (Monte Carlo)   sampling techniques\, thus imp
 lying overwhelming computational costs because   of the huge amount of que
 ries to the high-fidelity\, full-order coupled   PDE-ODEs model. To mitiga
 te this computational burden\, we replace the   high-fidelity model with c
 omputationally inexpensive projection-based   reduced-order models aimed a
 t reducing the state-space dimensionality. ROM   approximation errors on t
 he outputs of interest are finally taken into   account by means of statis
 tical error models built through Gaussian process   regression\, enhancing
  the accuracy of the whole UQ pipeline.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR