BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Open Vocabulary Confusion Networks for Speech Recognition - Keith 
 Vertanen
DTSTART:20070606T130000Z
DTEND:20070606T140000Z
UID:TALK7467@talks.cam.ac.uk
CONTACT:Christian Steinruecken
DESCRIPTION:A limitation of most current speech recognizers is that they o
 nly recognize words from a fixed vocabulary. If the user utters a novel wo
 rd\, the recognizer can only indicate it doesn't know\, or choose the "lea
 st-worst" word in its vocabulary. In this paper\, I explore a technique fo
 r addressing this deficiency using automatically derived units made up of 
 letters and phones. I show how these units can be merged to form likely ne
 w words. I further give an algorithm for doing this merger while maintaini
 ng a word lattice structure. This allows construction of a word confusion 
 network containing both in- and out-of-vocabulary (OOV) words. Experiments
  show these open vocabulary confusion networks provide improved word recog
 nition rates of up to 24% relative.\n\nFurther details can be found in the
  "paper":http://www.keithv.com/pub/openvocab/ or in chapter 6 of my "thesi
 s":http://www.keithv.com/pub/phd/
LOCATION:TCM Seminar Room\, Cavendish Laboratory\, Department of Physics
END:VEVENT
END:VCALENDAR