target word 0.00568911
source word 0.00533533
head word 0.005222452
english word 0.005000519
word order 0.00479144
large word 0.004773971
translation model 0.00473131
word baseline 0.004699284
word accuracy 0.0046490640000000005
primary word 0.004544007
mapping word 0.004536228
secondary word 0.0045142590000000005
word frequency 0.0045139640000000005
word subsequences 0.004510153
spanish word 0.004501875
middle word 0.004495715
alignment model 0.00393784
model alignment 0.00393784
translation models 0.003159955
speech translation 0.002986458
state translation 0.0029789779999999997
translation system 0.002957534
system translation 0.002957534
same translation 0.002934988
machine translation 0.0028952279999999997
translation string 0.0028923679999999998
state model 0.002883128
human translation 0.0028403049999999996
translation performance 0.0028298779999999996
translation accuracy 0.0028212339999999997
single translation 0.002798756
lation model 0.002783969
model size 0.002772115
reference translation 0.002747721
translation quality 0.0027179369999999997
translation bureau 0.002684425
ducer translation 0.002676717
spanish translation 0.002674045
builder translation 0.0026720499999999996
chine translation 0.00267061
translation lnodels 0.0026681559999999997
target words 0.00266039
transducer model 0.002659178
alignment search 0.00264825
transduction model 0.002619344
model con 0.002619033
model construction 0.0026139600000000002
ducer model 0.002580867
trace model 0.002578188
duction model 0.002576946
model builder 0.0025762
race model 0.002574447
target head 0.002428742
translation 0.00241358
alignment function 0.002385983
target language 0.002372041
model 0.00231773
source words 0.0023066099999999997
head words 0.002193732
source sentence 0.0020870899999999998
source language 0.002018261
target string 0.0019264879999999999
alignment func 0.001906482
alignment stage 0.001891145
builder alignment 0.00187858
target output 0.001851728
training corpus 0.001777617
target strings 0.0017667859999999998
training data 0.0017534809999999999
target partitions 0.001718018
lexical head 0.0017178999999999999
target subsequences 0.0017164429999999998
compound target 0.001714806
target subsequence 0.001707251
ranking target 0.0017028759999999999
target subse 0.0017028059999999999
phrase head 0.0016693239999999998
dependent words 0.0016622759999999999
training method 0.001630688
alignment 0.00162011
same source 0.001615328
translations head 0.0016099169999999998
different head 0.0016012079999999998
single words 0.0015978659999999999
source string 0.001572708
source length 0.001549532
cost function 0.001544545
statistical head 0.0015429149999999997
training algorithm 0.001535976
tion training 0.001520251
primary words 0.001515287
search algorithm 0.0015044160000000002
tion search 0.001488691
source sequence 0.0014617459999999999
source text 0.001452929
sentence length 0.001448782
target 0.0014477
training trans 0.001436315
training procedure 0.001426783
ing source 0.001425521
