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:Deep Neural Networks and Multigrid Methods - Jinch
ao Xu (Pennsylvania State University)
DTSTART;TZID=Europe/London:20191030T140500
DTEND;TZID=Europe/London:20191030T150500
UID:TALK134185AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/134185
DESCRIPTION:In this talk\, I will first give an introduction t
o several models and algorithms from two different
fields: (1) machine learning\, including logistic
regression\, support vector machine and deep neur
al networks\, and (2) numerical PDEs\, including f
inite element and multigrid methods. \; I will
then explore mathematical relationships between t
hese models and algorithms and demonstrate how suc
h relationships can be used to understand\, study
and improve the model structures\, mathematical pr
operties and relevant training algorithms for deep
neural networks. In particular\, I will demonstra
te how a new convolutional neural network known as
MgNet\, can be derived by making very minor modif
ications of a classic geometric multigrid method f
or the Poisson equation and then explore the theor
etical and practical potentials of MgNet.

LOCATION:Seminar Room 2\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR