BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Isaac Newton Institute Seminar Series
SUMMARY:Aspects of adaptive Galerkin FE for stochastic dir
ect and inverse problems - Martin Eigel (Weierstr
aß-Institut für Angewandte Analysis und Stochastik
)
DTSTART;TZID=Europe/London:20180207T090000
DTEND;TZID=Europe/London:20180207T100000
UID:TALK100102AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/100102
DESCRIPTION:Co-authors: Max Pfeffer (MPI MIS Leipzig)\,
Manuel Marschall (WIAS Berlin)\, Reinhold Schnei
der (TU Berlin)
The S
tochastic Galerkin Finite Element Method (SGFEM) i
s a common approach to numerically solve random PD
Es with the aim to obtain a functional representat
ion of the stochastic solution. As with any spectr
al method\, the curse of dimensionality renders th
e approach challenging when the randomness depends
on a large or countable infinite set of parameter
s. This makes function space adaptation and model
reduction strategies a necessity. We review adapti
ve SGFEM based on reliable a posteriori error esti
mators for affine and non-affine parameter represe
ntations. Based on this\, an adaptive explicit sam
pling-free Bayesian inversion in hierarchical tens
or formats can be derived. As an outlook onto curr
ent research\, a statistical learning viewpoint is
presented\, which connects concepts of UQ and mac
hine learning from a Variational Monte Carlo persp
ective.
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR