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:Applied and Computational Analysis
SUMMARY:R-adaptivity\, deep learning and the deep ritz met
hod - Chris Budd (University of Bath)
DTSTART;TZID=Europe/London:20231019T150000
DTEND;TZID=Europe/London:20231019T160000
UID:TALK207412AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/207412
DESCRIPTION:PINNS (physics informed neural nets) are becoming
increasingly popular methods for using deep learni
ng techniques to solve a wide variety of different
ial equations. They have been advertised as 'mesh
free methods' which can out perform traditional me
thods. But how good are they in practice? In this
talk I will look at how they compare with traditio
nal techniques such as the finite element method o
n different types of PDE\, linking their performan
ce to that of general nonlinear approximation met
hods such as Free Knot Splines. I will show that a
combination of 'traditional' numerical analysis a
nd deep learning can yield good results. But ther
e is still a lot to be learned about the performan
ce and reliability of a PINN based method.\n\nJoin
t work with Simone Appela\, Teo Deveney\, and Lisa
Kreusser.
LOCATION:Centre for Mathematical Sciences\, MR14
CONTACT:Nicolas Boulle
END:VEVENT
END:VCALENDAR