phrase translation 0.00616383
translation model 0.00546971
translation system 0.0041605159999999995
translation models 0.004111597
translation table 0.004067837
same translation 0.003910989
translation systems 0.003904729
training data 0.00387995
translation task 0.003872444
machine translation 0.00383952
translation results 0.003815422
english translation 0.003810088
translation quality 0.003743144
reference translation 0.003726524
measure translation 0.003718591
translation hypothesis 0.003701971
language data 0.00365514
chine translation 0.003645999
translation direction 0.003643386
translation errors 0.003641582
phrase table 0.003494387
target phrase 0.003474075
several phrase 0.003420666
phrase pair 0.0034128179999999998
translation 0.00336864
full phrase 0.003343351
new phrase 0.00333749
test data 0.003292902
language model 0.00327795
europarl phrase 0.003233179
phrase pairs 0.0032202339999999998
standard phrase 0.003204905
phrase tables 0.0031892789999999997
additional phrase 0.003182998
quality phrase 0.003169694
bilingual data 0.0031078729999999997
phrase posterior 0.003099827
phrase pos 0.0030860419999999998
alignment data 0.003077045
merged phrase 0.003074449
nal phrase 0.003067699
data set 0.003060623
nist data 0.003041515
development data 0.003030278
new data 0.00302056
monolingual data 0.00295707
english data 0.002919708
europarl data 0.002916249
news data 0.002898407
probability model 0.002894251
model probability 0.002894251
data sets 0.0028740709999999997
ing data 0.002871416
additional data 0.002866068
labeled data 0.002819589
data track 0.0027954729999999997
phrase 0.00279519
particular data 0.002779068
unlabeled data 0.0027676949999999997
newsgroup data 0.002751107
sparse data 0.002751107
model weights 0.002699848
model score 0.002623999
model scores 0.002528365
additional model 0.002488878
generative model 0.002466798
ibm model 0.002451517
mixture model 0.002446621
discriminative model 0.002435299
trigram model 0.002423272
model parameters 0.002410874
guage model 0.002410681
training sentence 0.002401178
current model 0.002390951
model idea 0.002387414
distortion model 0.002379976
model 0.00210107
language sentences 0.0020838059999999997
bilingual training 0.002031303
word alignment 0.0019492049999999999
test corpus 0.001900332
target language 0.0018557649999999999
training step 0.001827556
test sentence 0.00181413
language pair 0.0017945079999999998
additional training 0.001789498
word error 0.001785086
available training 0.001779713
word alignments 0.0017624
tional training 0.001756442
typical training 0.001730972
source language 0.001725202
training sen 0.001723844
different models 0.001721473
independent word 0.001701923
target phrases 0.0016938349999999999
target sentence 0.001678373
selection method 0.0016772059999999999
word penalty 0.00165552
word posterior 0.001655057
