BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Incremental adaptation of speech recognition based on macroscopic 
 time evolution system - Shinji Watanabe (NTT Communication Science Laborat
 ories\, Japan)
DTSTART:20110530T120000Z
DTEND:20110530T130000Z
UID:TALK31506@talks.cam.ac.uk
CONTACT:Kai Yu
DESCRIPTION:In a real environment\, acoustic and language features often v
 ary depending on the speakers\, speaking styles and topic changes. To acco
 mmodate these changes\, speech recognition approaches have to include the 
 incremental tracking of changing environments. In this talk\, I introduce 
 acoustic model adaptation based on macroscopic time evolution system\, whe
 re posterior distributions of acoustic model parameters are successively u
 pdated based on a macroscopic time scale. I also introduce a topic trackin
 g language model that can adaptively track changes in topics based on curr
 ent text information and previously estimated topic models in an on-line m
 anner. In addition\, I briefly introduce recent NTT CS Labs' speech recogn
 ition research\, which mainly focuses on meeting recognition.\n
LOCATION:Cambridge University Engineering Department\, Lecture Room 6
END:VEVENT
END:VCALENDAR