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:Uncertainty Quantification
SUMMARY:Multivariate polynomial quadrature via feature sel
 ection - Department of Mathematics\, University of
  Utah
DTSTART;TZID=Europe/London:20180713T170000
DTEND;TZID=Europe/London:20180713T180000
UID:TALK107920AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/107920
DESCRIPTION:Numerical quadrature rules that use point values a
 re ubiquitous tools for approximating integrals. S
 ome of the most popular rules achieve accuracy by 
 enforcing exactness for integrands in a finite-dim
 ensional polynomial space. When the integration do
 main is one-dimensional\, classical rules are avai
 lable and plentiful. In multidimensional domains w
 ith non-standard polynomial spaces and weights\, t
 he situation is far more complicated.\n\nWe will p
 resent a general methodology for numerically gener
 ating approximate polynomial quadrature rules in m
 ultidimensional situations. Our technique is based
  on recent advances in feature selection algorithm
 s\, allowing rigorous guarantees on feasibility an
 d computability of the selection problem. We show 
 that this approach allows one to generate excellen
 t quadrature rules in an automated way\, and demon
 strate the accuracy of these rules in applications
 .\n
LOCATION:EDC Loft meeting room (Inglis Building\, CUED)
CONTACT:Pranay Seshadri
END:VEVENT
END:VCALENDAR