C04-1033 |
procedure is the counterpart of the
|
training procedure
|
. Given a testing docu - ment
|
A00-2030 |
used the following multi-step
|
training procedure
|
which exploited the Penn TREEBANK
|
C04-1006 |
we will now sketch the standard
|
training procedure
|
for the lexicon model . The EM
|
C04-1006 |
translation probabilities . The standard
|
training procedure
|
of the statistical models uses
|
A94-1030 |
bilingual corpus . The unsupervised
|
training procedure
|
is described in detail in Fung
|
C00-2098 |
top of the parser stack . The
|
training procedure
|
of our probabilistic parser is
|
A97-1004 |
invalid sentence boundary . The
|
training procedure
|
requires no hand-crafted rules
|
C04-1045 |
) P ~ f Chier ( ~ f ; e ) The
|
training procedure
|
for the other model parameters
|
C04-1080 |
art results without the lengthy
|
training procedure
|
involved in other highperforming
|
C04-1060 |
in one language is given to the
|
training procedure
|
. It is important to note , however
|
C00-2141 |
local context templates made the
|
training procedure
|
very easy . * The three-stage
|
A83-1031 |
testing has begun using the same
|
training procedure
|
. It is too early to report results
|
C04-1032 |
computed as the result of the
|
training procedure
|
. In the source-totarget translation
|
A97-1051 |
subsequent manual or automatic
|
training procedures
|
. However , much of the drudgery
|
C04-1033 |
NPk = arg maxNP i2Ck jSNPij 3.2
|
Training procedure
|
Given an annotated training document
|
C00-2141 |
very easy . * The three-stage
|
training procedure
|
guarantees that only the useful
|
A83-1031 |
has not yet been possible . The
|
training procedure
|
has two parts . The first part
|
C04-1045 |
2001 ) for the conventional EM
|
training procedure
|
. Experimental results are reported
|
C04-1060 |
estimated with complete EM , while the
|
training procedure
|
for the IBM models samples from
|
C04-1032 |
in a fast and robust alignment
|
training procedure
|
. We also tested the more simple
|