BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Speech Recognition During Dictation Corrections - Keith Vertanen\,
  Inference Group
DTSTART:20060301T150000Z
DTEND:20060301T160000Z
UID:TALK4809@talks.cam.ac.uk
CONTACT:Phil Cowans
DESCRIPTION:In this talk\, I will describe an experiment in which 24 subje
 cts used a speech dictation style interface.  The behavior of subjects in 
 the face of errors will be described and their speaking style analyzed.  R
 ecognition performance on the subjects speech will be given using Microsof
 t\, Dragon and HTK speech recognizers.  The performance differences betwee
 n normal and hyperarticulated speech will be compared.  Early efforts to i
 mprove a recognizers acoustic model for spoken corrections will be describ
 ed.\n\nFor more details\, see the "paper":http://www.keithv.com/pub/speech
 during/ or chapter 3 of my "thesis":http://www.keithv.com/pub/phd/
LOCATION:Ryle Seminar Room\, Cavendish Laboratory
END:VEVENT
END:VCALENDAR