Sixth International Conference on Spoken Language Processing (ICSLP 2000)
October 16-20, 2000
Generalizing Prosodic Prediction of Speech Recognition Errors
Julia Hirschberg (1), Diane Litman (1), Marc Swerts (2)
(1) AT&T Labs - Research, Florham Park, NJ, USA
Since users of spoken dialogue systems have difficulty correcting
system misconceptions, it is important for automatic speech recognition
(ASR) systems to know when their best hypothesis is incorrect.
We compare results of previous experiments which showed
that prosody improves the detection of ASR errors to experiments
with a new system and new domain, the W99 conference registration
system. Our new results again show that prosodic features can
improve prediction of ASR misrecognitions over the use of other
standard techniques for ASR rejection.
(2) IPO, Eindhoven, The Netherlands, and CNTS, Antwerp, Belgium
Hirschberg, Julia / Litman, Diane / Swerts, Marc (2000):
"Generalizing prosodic prediction of speech recognition errors",
In ICSLP-2000, vol.1, 254-257.