Um, One
Large
Pizza: A Preliminary Study of
Disfluency Modelling for Improving ASR
|
Speaker: |
Ben Hutchinson and Cécile
Pereira |
|
|
SaySo!
(formerly Syrinx Speech Systems) |
|
Date: |
Wednesday,
4th July 2001 |
|
Time: |
11:00--12:30 |
|
Place: |
E6A
357, Macquarie University |
Abstract:
A corpus of spontaneous telephone transactions between
call
centre operators of a pizza company and its customers is examined for
disfluencies (fillers and speech repairs) with the aim of improving
automatic
speech recognition. From this, a subset of the customer orders is
selected as a
test set. An architecture is
presented
which allows filled pauses and repairs to be detected and
corrected. A language repair module removes
fillers and
reparanda and transforms utterances containing them into fluent
utterances. An
experiment on filled pauses using this module and architecture is then
described. A speech recognition grammar for
recognising
fluent speech is used to provide a baseline.
This grammar is then enriched with filled pauses, based on
their
placement in relation to syntactic boundaries. Evaluation is done at
the level
of understanding, using a metric on feature structures. Initial results indicate that
incorporating
filled pauses at syntactic
boundaries
improves the recognition results for spontaneous continuous speech
containing
disfluencies.
Bio:
Ben Hutchinson completed honours in linguistics at
Sydney
University in 1999, with a thesis on the formal learnability of
Optimality
Theory grammars. In 2000 he was a research assistant in the LTG
at
Macquarie University before joining SaySo!. Ben is currently working
on the
processing of spoken language, including the processing of language
disfluencies.
Cécile Pereira did her PhD thesis at the Speech, Hearing, and Language Research Centre, here at Macquarie University. The subject of her research was the Perception and Expression of Emotion in Speech. At SaySo! Cécile has been working with Ben on the recognition of language disfluencies, and amongst other things, is also involved in a SPIRT project with Macquarie University and UTS, investigating prosody in telephone service encounters.
Enquiries: sals@mri.mq.edu.au
|
Last modified: 12th June 2001 |