BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Disfluency detection in spoken learner English - Andrew Caines\, D
 TAL\, University of Cambridge
DTSTART:20150501T113000Z
DTEND:20150501T120000Z
UID:TALK59290@talks.cam.ac.uk
CONTACT:Tamara Polajnar
DESCRIPTION:Due to the non-canonical nature of spoken language (containing
  filled pauses\, non-standard grammatical variations\, hesitations and oth
 er disfluencies) and compounded by a lack of available training data\, spo
 ken language parsing has been a challenge for standard NLP tools. Recently
  the Redshift parser (Honnibal et al.\, CoNLL 2013) has been shown to be s
 uccessful in identifying grammatical relations and certain disfluencies in
  native speaker spoken language\, returning unlabelled dependency accuracy
  of 90.5% and a disfluency F-measure of 84.1% (Honnibal & Johnson\, TACL 2
 014). We investigate how this parser handles spoken data from learners of 
 English at various proficiency levels. Firstly\, we find that Redshift's p
 arsing accuracy on non-native speech data is comparable to Honnibal & John
 son's results\, with 91.1% of dependency relations correctly identified. H
 owever\, disfluency detection is markedly down\, with an F-measure of just
  47.8%. We consider why this should be\, and relate our findings to the us
 e of NLP technology for automatic language assessment and computer-assiste
 d language learning applications.\n
LOCATION:FW26\, Computer Laboratory
END:VEVENT
END:VCALENDAR