BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Virtual BSU Seminar: 'Confidence intervals for policy evaluation i
 n adaptive experiments’ - Dr Vitor Hadad\, Stanford University
DTSTART:20210527T130000Z
DTEND:20210527T140000Z
UID:TALK160543@talks.cam.ac.uk
CONTACT:Alison Quenault
DESCRIPTION:Randomized controlled trials are central to the scientific pro
 cess\, but they can be costly. For example\, a clinical trial may assign p
 atients to treatments that are detrimental to them. Adaptive experimental 
 designs\, such as multiarmed bandit algorithms\, reduce costs by increasin
 g the probability of assigning promising treatments over the course of the
  experiment. However\, because observations collected by these methods are
  dependent and their distribution is nonstationary\, statistical inference
  can be challenging. We propose a treatment-effect estimator that has an a
 symptotically unbiased and normal test statistic under straightforward\, r
 elatively weak conditions on the adaptive design. This estimator generaliz
 es for a variety of parameters of interest.
LOCATION:Virtual Seminar 
END:VEVENT
END:VCALENDAR