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:Probability
SUMMARY:Revisiting optimal scaling of Metropolis-Hastings
methods - Jure Vogrinc
DTSTART;TZID=Europe/London:20210511T140000
DTEND;TZID=Europe/London:20210511T150000
UID:TALK160138AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/160138
DESCRIPTION:Deriving useful bounds on the mixing of high dimen
sional Markov chain Monte Carlo (MCMC) algorithms
still seems impossibly difficult. In the analysis
of their performance we therefore restrict ourselv
es to answering simpler proxy questions\, such as
how to optimally\nchoose design parameters in an M
CMC method. We will review some classical results
on how to optimally scale the proposal variance in
the class of Metropolis-Hastings algorithms. Then
we will approach the same problem with a novel ax
iomatic framework enabling a more\ndetailed analys
is. We will look at counterexamples exhibiting ano
malous optimal scaling rates and identifying exact
smoothness assumptions required for classical sca
ling. We will discuss mathematical background\, in
tuitive understanding and further applications of
the\naxiomatic framework. \n\nThe talk is based on
joint work with Wilfrid Kendall.
LOCATION:Zoom
CONTACT:Perla Sousi
END:VEVENT
END:VCALENDAR