english data 0.00383146
parallel data 0.0035004500000000004
translation model 0.0034969199999999997
training data 0.00345765
english translation 0.00340283
different word 0.003117582
word alignment 0.0030468749999999997
data table 0.003037714
language model 0.00294898
new word 0.0029456559999999996
monolingual data 0.002787092
malagasy data 0.002743057
word substitution 0.0027398099999999996
joint word 0.002713076
available data 0.002710091
comparable data 0.002688797
annotated data 0.002683794
translation system 0.002650657
tuning data 0.002650134
lingual data 0.002648765
data types 0.002642314
lel data 0.0026289250000000003
allel data 0.0026283260000000003
word align 0.0026274849999999997
same translation 0.002626125
gual data 0.002623853
testing data 0.002623555
word order 0.0026221729999999998
hmm word 0.0026164599999999997
translation table 0.002609084
word vector 0.00260801
word level 0.0025985929999999997
word types 0.002587814
ferent word 0.0025771559999999997
word penalty 0.002572761
new translation 0.002571526
word orders 0.0025699679999999997
translation probabilities 0.002516179
english words 0.002474971
machine translation 0.002461348
ing translation 0.002426032
translation experiments 0.002407889
translation rules 0.0024073239999999997
translation pairs 0.002402902
only translation 0.002362556
alignment model 0.002339785
model probability 0.002288609
translation lexicon 0.002268888
english sentence 0.002253241
translation lexicons 0.002237972
seed translation 0.00222762
english translations 0.002226727
inverse translation 0.002221925
translation proba 0.00220315
translation lexi 0.0022009209999999998
chine translation 0.002197106
didate translation 0.002193408
english decipherment 0.002191302
parallel corpus 0.0021533940000000003
model score 0.002090902
language decipherment 0.0020704119999999998
english dependency 0.002068407
distortion model 0.002019995
channel model 0.00195126
dependency language 0.001947517
translation 0.00191494
ment model 0.001910691
guage model 0.0019061199999999999
english part 0.001905165
english gigaword 0.001900562
malagasy english 0.001887377
model prob 0.0018723379999999999
english train 0.001846197
english vocabulary 0.001809458
english penn 0.001793436
news test 0.001776433
english dependen 0.0017667949999999998
language malagasy 0.001766487
malagasy language 0.001766487
different test 0.001759475
parallel text 0.0017291150000000003
bigram language 0.001728542
language setting 0.0017158709999999999
different news 0.001673982
density language 0.0016648969999999998
gram language 0.001657963
tial language 0.001651727
ral language 0.0016512459999999999
official language 0.001645756
model 0.00158198
parallel part 0.001574155
sampling process 0.001565187
malagasy training 0.001513567
other models 0.001506482
non parallel 0.0014975790000000002
english 0.00148789
parallel cor 0.0014856090000000001
test set 0.001473162
rate training 0.0014443619999999998
extra training 0.001440588
