word alignment 0.0079318
alignment links 0.005947914
different alignment 0.005769602
human alignment 0.005525754
gold alignment 0.00541207
alignment precision 0.005392063
alignment errors 0.005380772
certain alignment 0.005368701
unsupervised alignment 0.0053684430000000005
alignment bleu 0.005367435
problematic alignment 0.0053674180000000005
supervised alignment 0.005338645
alignment schemes 0.005323469
man alignment 0.005320402
alignment accuracy 0.005291205
oracle alignment 0.005278818
alignment algo 0.005277178
alignment hurt 0.005275433
alignment 0.00497722
word translation 0.00463387
word corpus 0.00407702
english word 0.003714055
word pairs 0.003499074
chinese word 0.003447136
noisy word 0.003356567
certain word 0.003346061
word tokens 0.003344349
supervised word 0.003316005
com word 0.003300539
simple word 0.003298799
word corre 0.003297859
word trans 0.003284204
dependent word 0.0032648020000000002
word correspondences 0.00325634
regular word 0.00325288
discriminative word 0.003252788
improving word 0.003250828
translation model 0.00269842
translation models 0.002468431
translation rules 0.002181472
rule translation 0.002167003
machine translation 0.002165366
translation hypotheses 0.00210989
translation probabilities 0.002051405
incorrect translation 0.002035853
tree translation 0.0019975789999999998
bad translation 0.001976724
training corpus 0.001946207
english corpus 0.001881915
large corpus 0.00185269
english words 0.0018388150000000002
training data 0.001793596
other data 0.0017575310000000001
language model 0.001712701
translation 0.00167929
small corpus 0.0016173819999999999
linguistic data 0.001616874
human alignments 0.001580534
corpus size 0.001567448
large training 0.001554017
such links 0.001536103
parallel data 0.001526661
possible links 0.0014934850000000001
model scores 0.00149286
tion model 0.001467886
gold alignments 0.00146685
ldc corpus 0.0014551669999999998
ment data 0.001445769
aligned corpus 0.001427002
large language 0.001423821
unsupervised alignments 0.001423223
parallel training 0.0013805990000000001
small training 0.001318709
special links 0.0012925810000000001
human align 0.001281876
english text 0.0012780270000000002
cial links 0.001269778
regular links 0.001268994
training size 0.001268775
data items 0.001267423
data consortium 0.0012665649999999999
smt models 0.001239249
language pairs 0.001238065
sentence pair 0.001224355
training corpora 0.001222292
test set 0.001185481
limited training 0.001177177
such information 0.0011762930000000001
small set 0.001158296
example sentence 0.001153652
lation models 0.001129359
unsupervised align 0.0011245650000000001
corpus 0.00112244
training conditions 0.001120252
other studies 0.001107558
large number 0.001095064
other researchers 0.001092269
english side 0.001080597
words 0.00107934
large amount 0.001061412
