P03-1041 |
position in sentence ) within the
|
maximum approximation
|
. Depending on the direction
|
J87-1020 |
fixed point with respect to the
|
maximum approximation
|
. Comparison relations , combined
|
P03-1041 |
examining its consistency with the
|
maximum approximation
|
hypothesis by considering the
|
N04-1033 |
the preceding step , we used the
|
maximum approximation
|
for the sum over all segmentations
|
P03-1041 |
methods presented above used the
|
maximum approximation
|
to score candidate translation
|
D15-1123 |
constant . During search , the
|
maximum approximation
|
is applied rather than summing
|
P03-1041 |
pair imposes constraints over the
|
maximum approximation
|
hypothesis for sentences in which
|
C04-1032 |
than the ones obtained by the
|
maximum approximation
|
, i.e. by considering only the
|
E97-1037 |
likelihood criterion in the so-called
|
maximum approximation
|
. i.e. the likelihood criterion
|
E03-1055 |
following search criterion in
|
maximum approximation
|
which is used in combination
|
D12-1047 |
B ( S , Q ) Then , we use the
|
maximum approximation
|
to the sum : P ( Q | S ) Pz max
|
J03-1005 |
f1J | eI1 ) is computed in the
|
maximum approximation
|
using the distortion and the
|
J04-4002 |
together and performing search in
|
maximum approximation
|
, we obtain the following decision
|
P02-1038 |
convex . 1.3 Alignment Models and
|
Maximum Approximation
|
Typically , the probability Pr
|
J87-1020 |
ranges from intensional equality to
|
maximum approximation
|
and equivalence relations that
|
J07-1003 |
search is performed using the
|
maximum approximation
|
( see Equations ( 1 ) and ( 2
|
P02-1038 |
performed using the so-called
|
maximum approximation
|
: ˆeI1 = argmax eI 1 Pr
|
C00-2123 |
the inverted alignments in the
|
maximum approximation
|
, we obtain as search criterion
|
N10-1073 |
running time by specifiying a
|
maximum approximation
|
error bound . We employ the wellknown
|
P03-1041 |
model will be referred to as the
|
maximum approximation
|
, which under HMM alignment (
|