different word 0.004235855
word error 0.004003255
word posterior 0.003956014
model features 0.00387863
word pos 0.003848794
word level 0.003822877
word confidence 0.003793368
current word 0.003748777
discrete word 0.003725248
word lattices 0.003669944
word poste 0.003665157
word posi 0.003665157
model feature 0.0034235100000000003
semantic features 0.003014092
lexical features 0.00297206
probability features 0.00295686
syntactic features 0.002945014
linguistic features 0.0027041
additional features 0.002696291
confidence features 0.002680778
combination features 0.002636852
various features 0.002610008
bility features 0.002592675
features cer 0.002591026
guistic features 0.00256044
dence features 0.002556159
used features 0.002552809
gate features 0.002552809
lexical feature 0.00251694
probability feature 0.00250174
syntactic feature 0.002489894
language model 0.002341861
features 0.00230707
feature vector 0.0023008819999999998
feature selection 0.002242699
linear model 0.002219484
feature link 0.002182384
tic feature 0.002180595
effective feature 0.002168472
single feature 0.00216456
various feature 0.0021548879999999998
feature example 0.002146587
feature weights 0.002142944
feature combinations 0.002136866
tactic feature 0.002109279
other words 0.002108023
feature pool 0.002100884
feature combina 0.002099514
robustness feature 0.002097943
binary model 0.002082718
entropy model 0.002042608
maxent model 0.002040414
detection model 0.001957538
smt model 0.001895121
bayes model 0.0018750260000000001
feature 0.00185195
sentences words 0.00183398
sification model 0.001819194
correct words 0.001811227
translation errors 0.001771313
single words 0.0017661200000000001
words total 0.0017511920000000002
tween words 0.001717278
glish words 0.001715975
translation hypothesis 0.0017148660000000002
translation system 0.001712172
mchinese words 0.001699944
translation task 0.001638806
data corpus 0.001589327
training data 0.001580864
model 0.00157156
news translation 0.0015439590000000001
translation table 0.001528833
test corpus 0.001522858
reference translation 0.0015217020000000002
machine translation 0.001514919
translation hypotheses 0.001492398
translation systems 0.0014800590000000001
translation chinese 0.0014769590000000001
phrase translation 0.0014731800000000001
words 0.00145351
multiple translation 0.0014173940000000002
chine translation 0.0014132430000000002
translation quality 0.0014033510000000002
translation hypothe 0.0013756850000000002
translation hypoth 0.0013756850000000002
erated translation 0.001374203
evaluation data 0.001336002
test set 0.001309737
training set 0.0013041799999999998
target language 0.0013038
target sentence 0.001291256
maxent models 0.00123579
language pair 0.001212858
source sentence 0.00120707
posterior probability 0.001186144
information source 0.001178409
sentence level 0.001160974
corpus bleu 0.001157033
corpus sentences 0.001143352
