parallel data 0.00334775
training data 0.00316053
news translation 0.002908997
translation score 0.002835071
language model 0.00281441
translation probability 0.002741448
translation models 0.00271491
weibo translation 0.0026913370000000002
machine translation 0.002684226
translation task 0.002667331
translation quality 0.002574891
chinese translation 0.002559923
translation experiments 0.002507253
model training 0.0025015000000000003
extraction data 0.002496882
data extraction 0.002496882
parallel sentence 0.00248628
standard data 0.002472476
allel data 0.002462571
twitter data 0.0024537870000000002
translation problems 0.002449267
microblog data 0.0024171180000000002
data detection 0.002416722
annotated data 0.002414957
reasonable translation 0.0024146700000000003
croblog translation 0.0023988440000000002
data extrac 0.002394221
data attempts 0.002392187
translation tables 0.002391931
data yields 0.002390549
microblog translation 0.0023800880000000003
sentence translations 0.0023775899999999997
ferent translation 0.0023648750000000002
translation directions 0.00235459
translation examples 0.002354478
stantial translation 0.002353954
parallel training 0.00227884
word alignments 0.002179328
english language 0.002177903
alignment model 0.002127057
word alignment 0.0021231469999999997
language score 0.002116101
parallel sentences 0.002102994
translation 0.00207769
english sentence 0.0020724330000000003
parallel test 0.002036362
word error 0.001979613
target language 0.0019612839999999998
baseline model 0.0019128679999999999
arabic word 0.001883004
same sentence 0.00186765
different training 0.0018517070000000002
weibo parallel 0.0018466770000000001
chinese language 0.0018409529999999998
factor model 0.001839914
english words 0.0018305630000000001
ibm model 0.001825076
language pairs 0.0018240799999999998
language pair 0.0018164849999999998
reordering model 0.001798514
main language 0.0017935529999999998
viterbi model 0.0017811189999999998
viterbi word 0.001777209
various language 0.001763331
men model 0.001756084
model vogue 0.001755963
quality sentence 0.001750451
jargon word 0.001736548
word cday 0.00172851
word indexes 0.00172851
sentence pairs 0.00171861
parallel tweets 0.001713007
parallel corpora 0.001712454
sentence pair 0.001711015
correct language 0.001708458
training set 0.001700563
whole sentence 0.001675711
many sentence 0.001671303
language constraints 0.0016690109999999998
naive language 0.0016679
language detection 0.001660722
parallel span 0.001645468
language identification 0.001642984
language detector 0.001636112
news test 0.001634639
target words 0.001613944
good translations 0.001609634
parallel web 0.001595607
such translations 0.001589872
parallel segment 0.001581045
parallel segments 0.001580652
useful sentence 0.0015769550000000001
parallel messages 0.0015653
parallel candidates 0.001560216
sentence detection 0.001555252
example sentence 0.001548334
parallel dataset 0.001544498
parallel posts 0.001541459
lel sentence 0.001538124
mining parallel 0.001536857
