language model 0.00528732
translation model 0.00412619
model probability 0.003808434
model probabilities 0.0035866939999999996
first model 0.0034960209999999998
model scores 0.0034608539999999998
trigger model 0.0034518049999999996
model score 0.003426069
linear model 0.0034157579999999996
reordering model 0.003395796
latter model 0.003346735
model gains 0.003341316
guage model 0.003340923
model integration 0.003307249
model pbl 0.003295465
ond model 0.003295465
language models 0.003138589
model 0.00306189
target language 0.0027389800000000002
new language 0.0027156710000000002
such language 0.002689532
large language 0.002647346
language modeling 0.002608071
final language 0.002584854
forward language 0.0025581980000000002
target word 0.00251689
language mod 0.002489526
gram language 0.002482815
trained language 0.002480467
ward language 0.002478998
huge language 0.002471802
ferent language 0.002463013
language conditioning 0.002458918
backward language 0.002458918
dependency language 0.002458918
current word 0.0023700460000000002
triggered word 0.002275908
word choice 0.002268177
language 0.00222543
translation models 0.001977459
training data 0.00191009
source words 0.00187959
other words 0.00183067
target words 0.00173654
english words 0.001624684
test data 0.0016011419999999998
translation probabilities 0.001589104
translation results 0.001584385
correct words 0.0015634989999999999
phrase translation 0.001539482
boundary words 0.001523458
future words 0.0015024539999999999
chinese words 0.0014979849999999999
source sentence 0.001474424
translation decoder 0.001469976
english translation 0.001465994
translation hypothesis 0.001462794
italic words 0.00146079
stop words 0.00145914
tween words 0.001458456
adjacent words 0.001457489
speech recognition 0.001441876
machine translation 0.001429053
translation tasks 0.001410662
translation example 0.001398541
translation hypotheses 0.0013645060000000001
news data 0.001329378
translation quality 0.001322202
phrasal translation 0.001315045
same sentence 0.001312894
ical translation 0.00131273
reverse translation 0.0013101480000000001
speech recog 0.001308497
gual data 0.001228725
words 0.00122299
our training 0.00119647
guage models 0.001192192
training procedures 0.001171222
reference sentence 0.001168603
additional information 0.0011473149999999999
hanced models 0.001146849
test set 0.001144242
source phrase 0.001131782
sentence pairs 0.001109197
evaluation test 0.001072152
speech 0.00107205
translation 0.0010643
sentence length 0.00106393
sentence stepwise 0.001052257
ing process 0.001038771
lexicon rules 0.0010387670000000001
new system 0.001028484
information trigger 0.001027493
joint probability 0.001022321
other features 0.001016771
ing transduction 0.001014394
other scores 0.0010066440000000001
gram probability 0.001003929
new target 0.001003791
lexicon rule 9.98641E-4
