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:Approximate Bayesian Inference for Stochastic Proc
esses - Stumph\, M (Imperial College London)
DTSTART;TZID=Europe/London:20140425T095000
DTEND;TZID=Europe/London:20140425T102500
UID:TALK52188AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/52188
DESCRIPTION:Co-authors: Paul Kirk (Imperial College London)\,
Angelique Ale (Imperial College London)\, Ann Babt
ie (Imperial College London)\, Sarah Filippi (Impe
rial College London)\, Eszter Lakatos (Imperial Co
llege London)\, Daniel Silk (Imperial College Lond
on)\, Thomas Thorn (University of Edinburgh) \n\nW
e consider approximate Bayesian computation (ABC)
approaches to model the dynamics and evolution of
molecular networks. Initially conceived to cope wi
th problems with intractable likelihoods\, ABC has
gained popularity over the past decade. But there
are still considerable problems in applying ABC t
o real-world problems\, some of which are shared w
ith exact Bayesian inference\, but some are due to
the nature of ABC. Here we will present some rece
nt advances that allow us to apply ABC sequential
Monte Carlo (SMC) to real biological problems. The
rate of convergence of ABC-SMC depends crucially
on the schedule of thresholds\, ?t\, t=1\,2\,\,T\,
and the perturbation kernels used to generate pro
posals from the previous population of parameters.
We show how both of these can be tuned individual
ly\, and jointly. Careful calibration of the ABC-S
MC approach can result in a 10-fold reduction in t
he computational burden (or more). I will also pro
vide an overview of an alternative approach where\
, rather than approximating the likelihood in an A
BC framework\, we provide approximations to the ma
ster equation that describes the evolution of the
stochastic system\, that go beyond the conventiona
l linear noise approximation (LNA). This allows us
to tackle systems with ``interesting dynamics"\,
that are typically beyond the scope of the LNA\, a
nd we will show how to use such approaches in exac
t Bayesian inference procedures (including nested
sampling and SMC).\n
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:Mustapha Amrani
END:VEVENT
END:VCALENDAR