word alignment 0.00623041
alignment model 0.005004649999999999
word alignments 0.00475205
language word 0.004153683
translation model 0.00406111
alignment models 0.003945393
source word 0.0039326909999999994
alignment data 0.0038845869999999997
hmm word 0.003872648
target word 0.003833999
word align 0.0037807279999999997
statistical word 0.0037353609999999995
unsupervised word 0.0036763079999999997
hmm alignment 0.003625238
baseline word 0.003589897
automatic word 0.0035660329999999997
rare word 0.003545154
alignment methods 0.0035444269999999997
alignment error 0.0034500779999999997
alignment problem 0.0033999
alignment performance 0.003381249
single alignment 0.0033614469999999996
final alignment 0.0033440019999999996
supervised alignment 0.003340962
alignment accuracy 0.003319726
alignment points 0.0033126289999999997
likely alignment 0.003303043
annotated alignment 0.003299899
probable alignment 0.003295794
alignment preci 0.003288218
dicted alignment 0.003288218
translation models 0.003001853
alignment 0.0029915
translation table 0.0026001260000000003
machine translation 0.0025252580000000003
model figure 0.002507493
translation performance 0.002437709
markov model 0.002427101
translation quality 0.0023892920000000003
distortion model 0.00238484
translation systems 0.002381217
translation perfor 0.002372854
tional translation 0.0023580470000000003
translation sys 0.0023474340000000002
ment model 0.0023439760000000002
model uncertainty 0.002318
forward model 0.00231773
ward model 0.002311096
model bij 0.00230982
language words 0.0020724330000000003
translation 0.00204796
model 0.00201315
hidden alignments 0.001972364
rate alignments 0.001927151
new alignments 0.001885071
optional alignments 0.0018592880000000002
ble alignments 0.00183813
good alignments 0.00182561
sure alignments 0.001813403
alignments dif 0.0018110600000000002
many words 0.001777352
sentence corpus 0.001746871
training data 0.001698442
source sentence 0.0016853950000000001
bilingual sentence 0.001643477
source language 0.0016085539999999999
target sentence 0.001586703
ibm models 0.001562339
common words 0.001545907
other language 0.00154327
several words 0.001530886
alignments 0.00151314
target language 0.001509862
markov models 0.001367844
standard models 0.0013524280000000001
simple method 0.001346052
language pairs 0.00134547
unsupervised training 0.001242753
same form 0.00123153
such information 0.001198747
phrase base 0.0011964039999999999
new training 0.001177286
hmm align 0.001175556
posterior constraints 0.0011748219999999998
training cor 0.00116168
words 0.00115766
such constraints 0.001157197
baseline training 0.0011563419999999999
art phrase 0.001140448
training process 0.001134903
training scheme 0.001117133
source lan 0.001095496
posterior decoding 0.001093056
recent work 0.001077937
induction method 0.001056432
agreement constraints 0.00105515
additional information 0.001054422
maximization algorithm 0.001047493
many languages 0.0010382500000000001
search errors 0.001016327
