word alignment 0.0042502600000000005
word alignments 0.00413125
translation model 0.0039454
training data 0.00318557
translation models 0.0030680300000000002
language model 0.002894629
corpus model 0.002847161
alignment probability 0.002842094
alignment quality 0.002787271
alignment error 0.0027087829999999998
model probability 0.002696234
posterior alignment 0.0026772999999999996
alignment probabilities 0.002654741
word pair 0.00260199
machine translation 0.002550699
model parameters 0.002540138
translation quality 0.002521971
alignment accuracy 0.0025184829999999997
alignment statistics 0.002517128
annotated alignment 0.002496121
improved alignment 0.002478218
additional word 0.002468268
ble alignment 0.002466398
terior alignment 0.002466293
ibm translation 0.00246418
current model 0.0024362480000000002
manual word 0.002414526
model smoothing 0.002413218
hand alignments 0.0024038590000000004
manual alignments 0.002401816
word level 0.0023904020000000002
translation probabilities 0.002389441
example alignments 0.002381701
explicit word 0.0023815850000000003
level alignments 0.0023776920000000003
data set 0.002372891
training corpus 0.002372431
automatic alignments 0.0023691740000000004
sible alignments 0.002358128
plausible alignments 0.0023562690000000002
permissible alignments 0.002353926
probable alignments 0.00235385
ual alignments 0.0023481680000000003
bogus alignments 0.002346274
matic alignments 0.002346274
model parame 0.002344372
model probabil 0.002329199
training set 0.002302701
statistical translation 0.002280873
large data 0.002278315
translation mod 0.0022741370000000003
translation systems 0.002246731
probabilistic translation 0.002206299
sophisticated translation 0.002205475
new training 0.002204727
able translation 0.002200423
alignment 0.00217828
many training 0.002111653
ing data 0.002084302
alignments 0.00205927
sufficient data 0.002046063
model 0.00203242
training samples 0.002007595
data sets 0.001948062
verbmobil data 0.0019425549999999999
labeled data 0.001925513
testing data 0.0019245269999999999
default training 0.001922371
incomplete data 0.001914151
translation 0.00191298
entire training 0.001878711
training examples 0.001854324
training schemes 0.001848111
training material 0.001844025
same models 0.00177338
ibm models 0.0017062499999999999
tion models 0.0016152749999999998
test set 0.0016079
training 0.00155769
source sentence 0.001554222
new language 0.0015092460000000001
unmodified models 0.001496817
sentence pair 0.001470027
parallel corpus 0.00145987
many language 0.0014161719999999998
large set 0.0013954459999999998
sentence pairs 0.001381162
language pairs 0.001303354
ing corpus 0.001271163
example sentence 0.001262448
sentence level 0.001258439
previous work 0.0012528449999999998
log likelihood 0.001212655
test sets 0.001183071
posterior probability 0.001162834
models 0.00115505
eign language 0.001151009
uncommon language 0.00115023
verbmobil corpus 0.001129416
likelihood function 0.001127018
