BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:The Block Diagonal Infinite Hidden Markov Model - Tom Stepleton (C
 MU)
DTSTART:20090121T130000Z
DTEND:20090121T140000Z
UID:TALK16426@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:The Infinite Hidden Markov Model (IHMM) is a nonparametric Bay
 esian discrete time series model. By maintaining a posterior distribution 
 over transition and emission models for a countably infinite number of hid
 den states\, the IHMM flexibly accommodates data for which the "right" num
 ber of such states is not known.\n\nThe Block Diagonal Infinite Hidden Mar
 kov Model (BD-IHMM) extends the IHMM to target circumstances where the dat
 a alternate over time among a collection of distinct behavioral regimes\, 
 or "sub-behaviors". In ordinary HMMs\, alternation like this arises from t
 ransition matrices with nearly block-diagonal structure\, and the BD-IHMM'
 s prior induces similar forms in inferred dynamics\, here too with flexibi
 lity in the number of blocks. By inferring a hidden state sequence for the
  data\, the model also associates all parts of the data with a sub-behavio
 r\, a sequence-clustering capability that may also have useful analogs in 
 settings besides time series.\n\nThis talk will present the BD-IHMM\, desc
 ribe inference techniques for the model\, show recent results (these invol
 ve the Nintendo Wii and '90s USA alt-rock sensation Collective Soul)\, and
  discuss extensions and new directions under investigation now.
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR