A00-2041 |
parser 's coverage , AUTOSEM 's
|
repair
|
algorithm relies entirely on
|
A00-2028 |
for preventing , identifying and
|
repairing
|
problems that arise in the conversation
|
A00-2041 |
hand coded knowledge dedicated to
|
repair
|
. 1 Introduction In order for
|
A00-2041 |
operate both at parse time and
|
repair
|
time . The evaluation demonstrates
|
A00-2041 |
search for each sentence needing
|
repair
|
is derived from the set of partial
|
A00-2041 |
collected in English . For both
|
repair
|
approaches we used the meaning
|
A00-2028 |
their strategies for detecting and
|
repairing
|
problems that arise in conversation
|
A00-1046 |
error handling . The mean time to
|
repair
|
errors also was 4.3 times faster
|
A00-2041 |
fragmentary analysis is passed into the
|
repair
|
module . For each pair of vertices
|
A00-2041 |
and Lavie , to appear ) . The
|
repair
|
process begins as the genetic
|
A00-2041 |
fragmentary analyses passed on to the
|
repair
|
stage . An example ' is displayed
|
A00-2041 |
here demonstrates that AUTOSEM 's
|
repair
|
approach operates 200 times faster
|
A00-2041 |
robust parsers paired with separate
|
repair
|
modules , with separate knowledge
|
A00-2028 |
dialogue strategy in an attempt to
|
repair
|
the prob - lem . We can train
|
A00-2041 |
anal - ysis . Because the AUTOSEM
|
repair
|
algorithm runs significantly
|
A00-2041 |
functions can then be used in a
|
repair
|
stage to compose the fragments
|
A00-2041 |
only .33 seconds on average to
|
repair
|
a single parse while achieving
|
A00-2028 |
manager to modify its behavior to
|
repair
|
problems , and even perhaps ,
|
A00-2041 |
operate both at parse time and
|
repair
|
time . AUTOSEM is integrated
|
A00-2028 |
more sophisticated strategies to
|
repairing
|
problems , and even perhaps ,
|