BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Robust inference for intractable likelihood models using kernel di
 vergences - Francois-Xavier Briol (UCL)
DTSTART:20220211T130000Z
DTEND:20220211T140000Z
UID:TALK169907@talks.cam.ac.uk
CONTACT:Qingyuan Zhao
DESCRIPTION:Modern statistics and machine learning tools are being applied
  to increasingly complex phenomenon\, and as a result make use of increasi
 ngly complex models. A large class of such models are the so-called intrac
 table likelihood models\, where the likelihood is either too computational
  expensive to evaluate\, or impossible to write down in closed form. This 
 creates significant issues for classical approach such as maximum likeliho
 od estimation or Bayesian inference\, which are entirely reliant on evalua
 tions of a likelihood. In this talk\, we will cover several novel inferenc
 e schemes which by-pass this issue. These will be constructed from kernel-
 based discrepancies such as maximum mean discrepancies and kernel Stein di
 screpancies\, and can be used either in a frequentist or Bayesian framewor
 k. An important feature of our approach is that it will be provably robust
 \, in the sense that a small number of outliers or mild model misspecifica
 tion will not have a significant impact on parameter estimation. In partic
 ular\, we will show how the choice of kernel can allow us to trade statist
 ical efficiency with robustness. The methodology will then be illustrated 
 on a range of intractable likelihood models in signal processing and bioch
 emistry.
LOCATION:MR12\, Centre for Mathematical Sciences
END:VEVENT
END:VCALENDAR