#23889We describe a generative probabilistic model of natural language, which we call HBG, that takes advantage of detailed linguistic information to resolveambiguity.
other,6-3-H92-1026,ak
use a
<term>
corpus
</term>
of bracketed
<term>
sentences
</term>
, called a
<term>
Treebank
</term>
,
#23921We use a corpus of bracketedsentences, called a Treebank, in combination with decision tree building to tease out the relevant aspects of a parse tree that will determine the correct parse of a sentence.
lr,10-3-H92-1026,ak
bracketed
<term>
sentences
</term>
, called a
<term>
Treebank
</term>
, in combination with
<term>
decision
#23925We use a corpus of bracketed sentences, called aTreebank, in combination with decision tree building to tease out the relevant aspects of a parse tree that will determine the correct parse of a sentence.
other,36-3-H92-1026,ak
the correct
<term>
parse
</term>
of a
<term>
sentence
</term>
. This stands in contrast to the
#23951We use a corpus of bracketed sentences, called a Treebank, in combination with decision tree building to tease out the relevant aspects of a parse tree that will determine the correct parse of asentence.
other,33-3-H92-1026,ak
</term>
that will determine the correct
<term>
parse
</term>
of a
<term>
sentence
</term>
. This stands
#23948We use a corpus of bracketed sentences, called a Treebank, in combination with decision tree building to tease out the relevant aspects of a parse tree that will determine the correctparse of a sentence.
other,15-4-H92-1026,ak
further grammar tailoring via the usual
<term>
linguistic introspection
</term>
in the hope of generating the correct
#23968This stands in contrast to the usual approach of further grammar tailoring via the usuallinguistic introspection in the hope of generating the correct parse.
other,24-4-H92-1026,ak
the hope of generating the correct
<term>
parse
</term>
. In head-to-head tests against one
#23977This stands in contrast to the usual approach of further grammar tailoring via the usual linguistic introspection in the hope of generating the correctparse.
other,7-1-H92-1026,ak
generative probabilistic model
</term>
of
<term>
natural language
</term>
, which we call
<term>
HBG
</term>
,
#23872We describe a generative probabilistic model ofnatural language, which we call HBG, that takes advantage of detailed linguistic information to resolve ambiguity.
other,2-2-H92-1026,ak
</term>
.
<term>
HBG
</term>
incorporates
<term>
lexical , syntactic , semantic , and structural information
</term>
from the
<term>
parse tree
</term>
into
#23893HBG incorporateslexical , syntactic , semantic , and structural information from the parse tree into the disambiguation process in a novel way.
lr,3-3-H92-1026,ak
process
</term>
in a novel way . We use a
<term>
corpus
</term>
of bracketed
<term>
sentences
</term>
#23918We use acorpus of bracketed sentences, called a Treebank, in combination with decision tree building to tease out the relevant aspects of a parse tree that will determine the correct parse of a sentence.
other,20-1-H92-1026,ak
, that takes advantage of detailed
<term>
linguistic information
</term>
to resolve
<term>
ambiguity
</term>
.
#23885We describe a generative probabilistic model of natural language, which we call HBG, that takes advantage of detailedlinguistic information to resolve ambiguity.
measure(ment),28-5-H92-1026,ak
<term>
P-CFG
</term>
, increasing the
<term>
parsing accuracy rate
</term>
from 60 % to 75 % , a 37 % reduction
#24007In head-to-head tests against one of the best existing robust probabilistic parsing models, which we call P-CFG, the HBG model significantly outperforms P-CFG, increasing theparsing accuracy rate from 60% to 75%, a 37% reduction in error.
model,17-5-H92-1026,ak
parsing models
</term>
, which we call
<term>
P-CFG
</term>
, the
<term>
HBG model
</term>
significantly
#23996In head-to-head tests against one of the best existing robust probabilistic parsing models, which we callP-CFG, the HBG model significantly outperforms P-CFG, increasing the parsing accuracy rate from 60% to 75%, a 37% reduction in error.
model,13-1-H92-1026,ak
natural language
</term>
, which we call
<term>
HBG
</term>
, that takes advantage of detailed
#23878We describe a generative probabilistic model of natural language, which we callHBG, that takes advantage of detailed linguistic information to resolve ambiguity.
model,3-1-H92-1026,ak
OCR accuracy
</term>
. We describe a
<term>
generative probabilistic model
</term>
of
<term>
natural language
</term>
,
#23868We describe agenerative probabilistic model of natural language, which we call HBG, that takes advantage of detailed linguistic information to resolve ambiguity.
tech,15-3-H92-1026,ak
Treebank
</term>
, in combination with
<term>
decision tree building
</term>
to tease out the relevant aspects
#23930We use a corpus of bracketed sentences, called a Treebank, in combination withdecision tree building to tease out the relevant aspects of a parse tree that will determine the correct parse of a sentence.
model,20-5-H92-1026,ak
which we call
<term>
P-CFG
</term>
, the
<term>
HBG model
</term>
significantly outperforms
<term>
P-CFG
#23999In head-to-head tests against one of the best existing robust probabilistic parsing models, which we call P-CFG, theHBG model significantly outperforms P-CFG, increasing the parsing accuracy rate from 60% to 75%, a 37% reduction in error.
model,9-5-H92-1026,ak
tests against one of the best existing
<term>
robust probabilistic parsing models
</term>
, which we call
<term>
P-CFG
</term>
#23988In head-to-head tests against one of the best existingrobust probabilistic parsing models, which we call P-CFG, the HBG model significantly outperforms P-CFG, increasing the parsing accuracy rate from 60% to 75%, a 37% reduction in error.
other,26-3-H92-1026,ak
tease out the relevant aspects of a
<term>
parse tree
</term>
that will determine the correct
<term>
#23941We use a corpus of bracketed sentences, called a Treebank, in combination with decision tree building to tease out the relevant aspects of aparse tree that will determine the correct parse of a sentence.
other,17-2-H92-1026,ak
the
<term>
parse tree
</term>
into the
<term>
disambiguation process
</term>
in a novel way . We use a
<term>
corpus
#23908HBG incorporates lexical, syntactic, semantic, and structural information from the parse tree into thedisambiguation process in a novel way.