translation model 0.00510625
language model 0.0037739089999999998
lexicon model 0.003467199
standard model 0.003394258
model parameters 0.003345057
bilingual model 0.0032289239999999998
model estimation 0.003219375
model structure 0.003204179
tion model 0.003190329
backoff model 0.003182612
complex model 0.003178222
optimal model 0.003166481
model search 0.003164721
conditional model 0.0031464039999999998
particular model 0.003146132
current model 0.003117482
field model 0.003116692
translation models 0.003111187
bayesian model 0.003109219
explicit model 0.003103695
model redundancy 0.003095766
model selection 0.0030947039999999998
lation model 0.0030926589999999998
model complexity 0.0030764679999999997
discussion model 0.003075118
model struc 0.003071103
optimised model 0.003070886
model sizes 0.003068338
unwarranted model 0.003065726
robust model 0.003065631
translation lexicon 0.002890049
phrase translation 0.0028617810000000003
model 0.0028417
translation table 0.002747558
target word 0.00269955
machine translation 0.0026929420000000003
translation process 0.0026229680000000003
statistical translation 0.0026148910000000003
translation probabilities 0.0025623900000000003
translation quality 0.002522359
robust translation 0.0024884810000000003
sign translation 0.002487186
translation 0.00226455
source language 0.002154979
target language 0.001947429
unknown word 0.0019083289999999998
target words 0.001856365
different language 0.001848308
source lexicon 0.001848269
language models 0.0017788460000000002
monolingual features 0.001775038
different models 0.001762736
source languages 0.001651206
source types 0.001643111
target lexicon 0.0016407190000000001
source cluster 0.0016284749999999999
lexical source 0.0016143429999999999
binary features 0.0015875309999999999
exact source 0.001564111
source type 0.001555592
training data 0.001551813
target phrases 0.001524105
monolingual feature 0.001503435
welsh source 0.001499979
source tokens 0.001486213
other language 0.001475144
lexicon models 0.001472136
language pairs 0.001458976
same target 0.001456925
source lex 0.0014449979999999999
source vocabu 0.0014449979999999999
target languages 0.001443656
target types 0.0014355610000000001
target cluster 0.001420925
feature space 0.001406686
standard models 0.001399195
such models 0.001387449
alignment process 0.001386918
parallel data 0.0013853440000000002
similar feature 0.001357891
corresponding target 0.001352352
target type 0.001348042
language pair 0.001347539
parallel training 0.0013404649999999999
training corpus 0.0013057849999999998
target lexicons 0.00130052
target vocabulary 0.0012857480000000002
target tokens 0.001278663
target distributions 0.001278258
target lan 0.001277514
baseline models 0.001274954
edit feature 0.001267776
clustering target 0.001267764
features 0.0012651
target vocabularies 0.001237381
ibm models 0.001225807
source 0.00122277
certain language 0.001213284
training set 0.0012100219999999998
test data 0.001209646
