#22538We present an efficient algorithm for chart-based phrase structure parsing of natural language that is tailored to the problem of extracting specific information fromunrestricted texts where many of the words are unknown and much of the text is irrelevant to the task.
other,38-6-A92-1027,ak
number of
<term>
edges
</term>
, since only
<term>
edges
</term>
with a valid
<term>
semantic interpretation
#22712A further reduction in the search space is achieved by using semantic rather than syntactic categories on the terminal and non-terminal edges, thereby reducing the amount of ambiguity and thus the number of edges, since onlyedges with a valid semantic interpretation are ever introduced.
other,8-3-A92-1027,ak
new
<term>
edge
</term>
is added to the
<term>
chart
</term>
, the
<term>
algorithm
</term>
checks
#22579As each new edge is added to thechart, the algorithm checks only the topmost of the edges adjacent to it, rather than all such edges as in conventional treatments.
other,2-4-A92-1027,ak
conventional treatments . The resulting
<term>
spanning edges
</term>
are insured to be the correct ones
#22606The resultingspanning edges are insured to be the correct ones by carefully controlling the order in which edges are introduced so that every final constituent covers the longest possible span.
other,14-6-A92-1027,ak
achieved by using semantic rather than
<term>
syntactic categories
</term>
on the
<term>
terminal and non-terminal
#22688A further reduction in the search space is achieved by using semantic rather thansyntactic categories on the terminal and non-terminal edges, thereby reducing the amount of ambiguity and thus the number of edges, since only edges with a valid semantic interpretation are ever introduced.
other,18-4-A92-1027,ak
carefully controlling the order in which
<term>
edges
</term>
are introduced so that every final
#22622The resulting spanning edges are insured to be the correct ones by carefully controlling the order in whichedges are introduced so that every final constituent covers the longest possible span.
other,25-4-A92-1027,ak
are introduced so that every final
<term>
constituent
</term>
covers the longest possible span
#22629The resulting spanning edges are insured to be the correct ones by carefully controlling the order in which edges are introduced so that every finalconstituent covers the longest possible span.
other,42-6-A92-1027,ak
only
<term>
edges
</term>
with a valid
<term>
semantic interpretation
</term>
are ever introduced . In this paper
#22716A further reduction in the search space is achieved by using semantic rather than syntactic categories on the terminal and non-terminal edges, thereby reducing the amount of ambiguity and thus the number of edges, since only edges with a validsemantic interpretation are ever introduced.
other,28-6-A92-1027,ak
</term>
, thereby reducing the amount of
<term>
ambiguity
</term>
and thus the number of
<term>
edges
#22702A further reduction in the search space is achieved by using semantic rather than syntactic categories on the terminal and non-terminal edges, thereby reducing the amount ofambiguity and thus the number of edges, since only edges with a valid semantic interpretation are ever introduced.
other,10-2-A92-1027,ak
efficiency
</term>
through a reduction of its
<term>
search space
</term>
. As each new
<term>
edge
</term>
is
#22568The parser gains algorithmic efficiency through a reduction of itssearch space.
other,22-1-A92-1027,ak
the problem of extracting specific
<term>
information
</term>
from
<term>
unrestricted texts
</term>
#22536We present an efficient algorithm for chart-based phrase structure parsing of natural language that is tailored to the problem of extracting specificinformation from unrestricted texts where many of the words are unknown and much of the text is irrelevant to the task.
other,35-5-A92-1027,ak
be deduced despite the presence of
<term>
unknown words
</term>
. A further reduction in the
<term>
#22671This is facilitated through the use of phrase boundary heuristics based on the placement of function words, and by heuristic rules that permit certain kinds of phrases to be deduced despite the presence ofunknown words.
other,15-5-A92-1027,ak
heuristics
</term>
based on the placement of
<term>
function words
</term>
, and by
<term>
heuristic rules
</term>
#22651This is facilitated through the use of phrase boundary heuristics based on the placement offunction words, and by heuristic rules that permit certain kinds of phrases to be deduced despite the presence of unknown words.
other,11-1-A92-1027,ak
phrase structure parsing
</term>
of
<term>
natural language
</term>
that is tailored to the problem of
#22525We present an efficient algorithm for chart-based phrase structure parsing ofnatural language that is tailored to the problem of extracting specific information from unrestricted texts where many of the words are unknown and much of the text is irrelevant to the task.
tech,4-1-A92-1027,ak
presented . We present an efficient
<term>
algorithm
</term>
for
<term>
chart-based phrase structure
#22518We present an efficientalgorithm for chart-based phrase structure parsing of natural language that is tailored to the problem of extracting specific information from unrestricted texts where many of the words are unknown and much of the text is irrelevant to the task.
other,5-6-A92-1027,ak
</term>
. A further reduction in the
<term>
search space
</term>
is achieved by using semantic rather
#22679A further reduction in thesearch space is achieved by using semantic rather than syntactic categories on the terminal and non-terminal edges, thereby reducing the amount of ambiguity and thus the number of edges, since only edges with a valid semantic interpretation are ever introduced.
other,18-6-A92-1027,ak
<term>
syntactic categories
</term>
on the
<term>
terminal and non-terminal edges
</term>
, thereby reducing the amount of
<term>
#22692A further reduction in the search space is achieved by using semantic rather than syntactic categories on theterminal and non-terminal edges, thereby reducing the amount of ambiguity and thus the number of edges, since only edges with a valid semantic interpretation are ever introduced.
model,7-5-A92-1027,ak
is facilitated through the use of
<term>
phrase boundary heuristics
</term>
based on the placement of
<term>
function
#22643This is facilitated through the use ofphrase boundary heuristics based on the placement of function words, and by heuristic rules that permit certain kinds of phrases to be deduced despite the presence of unknown words.
other,18-3-A92-1027,ak
</term>
checks only the topmost of the
<term>
edges
</term>
adjacent to it , rather than all
#22589As each new edge is added to the chart, the algorithm checks only the topmost of theedges adjacent to it, rather than all such edges as in conventional treatments.
other,3-2-A92-1027,ak
task . The
<term>
parser
</term>
gains
<term>
algorithmic efficiency
</term>
through a reduction of its
<term>
search
#22561The parser gainsalgorithmic efficiency through a reduction of its search space.