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:Machine Learning @ CUED
SUMMARY:Structural Markov laws / Geometry and HMC - Dr Sim
on Byrne
DTSTART;TZID=Europe/London:20150609T110000
DTEND;TZID=Europe/London:20150609T120000
UID:TALK59712AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/59712
DESCRIPTION:This talk will focus on two particular aspects of
my research:\n\nSuppose that we wish to infer the
structure of a graphical model: how should we choo
se a prior over the space of possible graphs? I'll
introduce the notion of a structural Markov prope
rty\, which requires that the structure of distinc
t components of the graph be conditionally indepen
dent given the existence of a separating component
. This characterises an exponential family that is
conjugate under sampling from compatible Markov d
istributions.\n\nIn the second part\, I will talk
about various geometric aspects of the Hamiltonian
/Hybrid Monte Carlo (HMC) algorithm. I will explai
n how HMC can be extended to manifolds\, such as s
pheres and Stiefel manifolds\n(the manifold of ort
hogonal matrices). I will also describe how this g
eometric understanding can guide the optimal tunin
g of the algorithm.\n
LOCATION:Engineering Department\, CBL Room BE-438
CONTACT:
END:VEVENT
END:VCALENDAR