#23755The weighted error for the entirespoken language system on the same test set is 49.3%.
other,12-4-H92-1016,ak
language system
</term>
on the same
<term>
test set
</term>
is 49.3 % . Similar results were
#23761The weighted error for the entire spoken language system on the sametest set is 49.3%.
other,9-5-H92-1016,ak
also obtained on the February '92
<term>
benchmark evaluation
</term>
. This paper describes three relatively
#23776Similar results were also obtained on the February '92benchmark evaluation.
tool,7-1-H92-1016,ak
paper describes the status of the
<term>
MIT ATIS system
</term>
as of February 1992 , focusing especially
#23662This paper describes the status of theMIT ATIS system as of February 1992, focusing especially on the changes made to the SUMMIT recognizer.
lr,26-2-H92-1016,ak
</term>
, and refinements made to the
<term>
lexicon
</term>
. Together with the use of a larger
#23707These include context-dependent phonetic modelling, the use of a bigram language model in conjunction with a probabilistic LR parser, and refinements made to thelexicon.
tech,2-2-H92-1016,ak
SUMMIT recognizer
</term>
. These include
<term>
context-dependent phonetic modelling
</term>
, the use of a
<term>
bigram language
#23683These includecontext-dependent phonetic modelling, the use of a bigram language model in conjunction with a probabilistic LR parser, and refinements made to the lexicon.
tool,23-1-H92-1016,ak
especially on the changes made to the
<term>
SUMMIT recognizer
</term>
. These include
<term>
context-dependent
#23678This paper describes the status of the MIT ATIS system as of February 1992, focusing especially on the changes made to theSUMMIT recognizer.
lr,7-3-H92-1016,ak
Together with the use of a larger
<term>
training set
</term>
, these modifications combined to
#23716Together with the use of a largertraining set, these modifications combined to reduce the speech recognition word and sentence error rates by a factor of 2.5 and 1.6, respectively, on the October '91 test set.
measure(ment),1-4-H92-1016,ak
October '91
<term>
test set
</term>
. The
<term>
weighted error
</term>
for the entire
<term>
spoken language
#23750Theweighted error for the entire spoken language system on the same test set is 49.3%.
model,10-2-H92-1016,ak
phonetic modelling
</term>
, the use of a
<term>
bigram language model
</term>
in conjunction with a
<term>
probabilistic
#23691These include context-dependent phonetic modelling, the use of abigram language model in conjunction with a probabilistic LR parser, and refinements made to the lexicon.
measure(ment),16-3-H92-1016,ak
modifications combined to reduce the
<term>
speech recognition word and sentence error rates
</term>
by a factor of 2.5 and 1.6 , respectively
#23725Together with the use of a larger training set, these modifications combined to reduce thespeech recognition word and sentence error rates by a factor of 2.5 and 1.6, respectively, on the October '91 test set.
other,37-3-H92-1016,ak
respectively , on the October '91
<term>
test set
</term>
. The
<term>
weighted error
</term>
for
#23746Together with the use of a larger training set, these modifications combined to reduce the speech recognition word and sentence error rates by a factor of 2.5 and 1.6, respectively, on the October '91test set.