BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Theoretical guarantees in Bayesian non parametric hidden Markov mo
 dels - Elodie Vernet
DTSTART:20161130T160000Z
DTEND:20161130T170000Z
UID:TALK69076@talks.cam.ac.uk
CONTACT:41319
DESCRIPTION:Hidden Markov models (HMMs) have been widely used in diverse f
 ields such as speech recognition\, genomics or econometrics. Because param
 etric modelling of HMMs may lead to\npoor results in practice recent inter
 est in using non parametric HMMs appeared in applications. Yet little thou
 ghts have been given to theory in this framework.\nIn this talk I will fir
 st define and describe the model I'm interested in\, namely Bayesian non p
 arametric hidden Markov models. I will then present the problems I study i
 n these models (namely concerning the asymptotic behaviour of the posterio
 r distribution). I will finally give some theoretical guarantees I have ob
 tained in these models.\n
LOCATION:MR14\, Centre for Mathematical Sciences
END:VEVENT
END:VCALENDAR