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:Statistics
SUMMARY:Stochastic Causal Programming for Bounding Treatme
nt Effects - Ricardo Silva (UCL)
DTSTART;TZID=Europe/London:20230224T140000
DTEND;TZID=Europe/London:20230224T150000
UID:TALK194905AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/194905
DESCRIPTION:Causal effect estimation is important for many tas
ks in the natural and social sciences. We design a
lgorithms for the continuous partial identificatio
n problem: bounding the effects of multivariate\,
continuous treatments when unmeasured confounding
makes identification impossible. Specifically\, we
cast causal effects as objective functions within
a constrained optimization problem\, and minimize
/maximize these functions to obtain bounds. We com
bine flexible learning algorithms with Monte Carlo
methods to implement a family of solutions under
the name of stochastic causal programming. In part
icular\, we show how the generic framework can be
efficiently formulated in settings where auxiliary
variables are clustered into pre-treatment and po
st-treatment sets\, where no fine-grained causal g
raph can be formulated. Contrasted to other generi
c approaches\, this highly simplifies the problem
and has advantages concerning how to encode struct
ural knowledge without explicitly constructing lat
ent hidden common causes.\n \nJoint work with Kirt
an Padh\, Jakob Zeitler\, David Watson\, Matt Kusn
er and Niki Kilbertus.
LOCATION:MR12\, Centre for Mathematical Sciences
CONTACT:Qingyuan Zhao
END:VEVENT
END:VCALENDAR