BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CERN//INDICO//EN
BEGIN:VEVENT
SUMMARY:Reduced Order Surrogate Models for PDE-Constrained Optimization an
 d Inverse Problems [Hausdorff Colloquium]
DTSTART:20251119T141500Z
DTEND:20251119T154500Z
DTSTAMP:20260420T161300Z
UID:indico-event-635@math-events.uni-bonn.de
DESCRIPTION:Speakers: Mario Ohlberger (Universität Münster)\n\nAbstract:
 \nClassically\, model order reduction for parameterized systems is based o
 n a so-called offline phase\, where reduced approximation spaces are const
 ructed and the reduced parameterized system is built\, followed by an onli
 ne phase\, where the reduced system can be cheaply evaluated in a multi-qu
 ery context. In this contribution\, instead\, we follow an active learning
  or enrichment approach where a multi-fidelity hierarchy of reduced order 
 models is constructed on-the-fly while exploring a parameterized system. T
 o this end we focus on learning based reduction methods in the context of 
 PDE constrained optimization and inverse problems and evaluate their overa
 ll efficiency. We discuss learning strategies\, such as adaptive enrichmen
 t within a trust region optimization framework as well as a combination o
 f reduced order models with machine learning approaches. Concepts of rigor
 ous certification and convergence will be presented\, as well as numerical
  experiments that demonstrate the efficiency of the proposed approaches.\
 nWebsite of the Hausdorff Colloquium\n\nhttps://math-events.uni-bonn.de/ev
 ent/635/
LOCATION:Endenicher Allee 60/1-016 - Lipschitzsaal (Mathezentrum)
URL:https://math-events.uni-bonn.de/event/635/
END:VEVENT
END:VCALENDAR