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:Mordell Lectures
SUMMARY:Bayesian inference in infinite dimensions - Aad va
n der Vaart (Delft)
DTSTART;TZID=Europe/London:20230504T170000
DTEND;TZID=Europe/London:20230504T180000
UID:TALK184304AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/184304
DESCRIPTION:The Bayesian statistical method consists of updati
ng a prior probability distribution\nover the unkn
own parameters of a stochastic system into a poste
rior probability distribution \nafter seeing the s
ystem's output. It is perhaps the oldest statistic
al paradigm\, going\nback to the 18th century\, in
abstract terms as straightforward and elegant as
can be\, and \nwith the promise of not only giving
a best guess of the system parameters\, but also
a\nquantification of remaining uncertainty. Only i
n the last two decades has the method\nbeen applie
d to infinite-dimensional parameters\, most recent
ly to inverse problems\ndefined e.g. by PDEs or in
machine learning. We discuss some of the mathemat
ical\nissues\, with a main focus on the question w
hether the method works and when\, and\nhow we can
define "works". We review some classical success
stories and recent \nfindings and open questions\,
borrowing from our own work and that of others.\n
\nThe talk will be followed by a wine reception in
the Central Core CMS\n
LOCATION:MR2\, CMS
CONTACT:HoD Secretary\, DPMMS
END:VEVENT
END:VCALENDAR