BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:« Context-Aware Model Reduction »
DTSTART:20181030T161500
DTEND:20181030T171500
DTSTAMP:20260406T184048Z
UID:e8c5f03554d0a74c327049a73f8a0e4905bbb06d4d5d51745a30318a
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Benjamin Peherstorfer\n\nAbstract:\nTraditional model re
 duction speeds up computations by replacing expensive high-fidelity models
  with low-cost reduced models. Thus\, the goal of traditional model reduct
 ion is to construct reduced models that approximate well the high-fidelity
  models\; however\, reduced models are increasingly used together with hig
 h-fidelity models\, which means that the primary purpose of reduced models
  becomes supporting computations with the high-fidelity models rather than
  approximating and replacing them. In this presentation\, we propose conte
 xt-aware model reduction strategies that explicitly construct reduced mode
 ls for being used together with other models. In the first part\, we prese
 nt an approach to dynamically couple reduced models with the high-fidelity
  model\, where the reduced models are adapted in a context-aware sense wit
 h sparse data from the high-fidelity model. Our numerical examples demonst
 rate that the dynamic coupling is particularly beneficial in case of trans
 port-dominated problems\, where our context-aware approach achieves signif
 icant speedups\, whereas traditional reduced models are even more costly t
 o evaluate than the high-fidelity models. In the second part of the presen
 tation\, we introduce the adaptive multifidelity Monte Carlo (AMFMC) metho
 d that constructs reduced models that support the multifidelity estimation
  of statistics of high-fidelity model outputs. Our analysis shows that our
  context-aware reduced models optimally reduce the runtime of the multifid
 elity estimation\, even though the context-aware reduced models are less a
 ccurate in the sense of traditional model reduction.\n 
LOCATION:MA A3 30 https://plan.epfl.ch/?room=MAA330
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR