domain data 0.00325016
parallel data 0.00323836
training data 0.003074019
test data 0.002999692
target data 0.0029246719999999997
document data 0.002863461
translation model 0.00272692
source data 0.002688066
small data 0.002661125
data set 0.002645349
such data 0.0026424919999999998
translation system 0.002630146
comparable data 0.0025902909999999998
data our 0.002568149
phrase translation 0.002539979
subtitles data 0.002529787
unlabeled data 0.002512361
tuning data 0.002496869
data sets 0.002488709
data sparsity 0.002473043
actual data 0.002472887
parliament data 0.002468821
development data 0.002445408
parable data 0.002438152
translation probability 0.002232083
training words 0.002220199
word pair 0.002199176
word pairs 0.002150119
language model 0.00214039
oov word 0.002127934
translation systems 0.002103373
machine translation 0.002087533
translation probabilities 0.002083393
english words 0.002044039
word tokens 0.002000844
context word 0.001999142
translation error 0.001993026
translation perfor 0.001963566
main word 0.001912063
new features 0.001872346
word similarities 0.0018659430000000001
words news 0.001853767
word return 0.0018529100000000001
oov words 0.001852234
other language 0.00184582
new domain 0.0018424959999999999
domain corpus 0.0018234140000000002
target domain 0.001819312
parallel corpus 0.001811614
parallel target 0.001807512
unseen words 0.00170908
translation 0.00169988
unknown words 0.001677847
subtitles words 0.001675967
language pairs 0.001663829
emea words 0.001650813
language models 0.001649906
good language 0.001631647
frequent words 0.001607164
news domain 0.001602227
german words 0.001600533
words sents 0.001597806
foreign words 0.001595427
vocabulary words 0.0015938179999999999
source domain 0.0015827060000000001
man words 0.001577323
context features 0.0015017519999999999
new translations 0.001472604
phrase table 0.001469607
domain adaptation 0.001454174
phrase pair 0.0014396349999999999
indicator features 0.001401944
phrase pairs 0.001390578
different target 0.001369757
new pair 0.001369632
orthographic features 0.001367222
parliament domain 0.001363461
form domain 0.001353006
same english 0.001344229
english text 0.001337356
plentiful parallel 0.0013280850000000001
challenges domain 0.001324976
domain divergence 0.001324976
words 0.00132394
first feature 0.001277763
target task 0.001255546
oracle system 0.001243469
phrase probabilities 0.001223612
new domains 0.00121578
tion system 0.001210317
lation system 0.001209631
machine learning 0.001204907
original phrase 0.001199171
target domains 0.001192596
such translations 0.00116724
baseline performance 0.001165097
training samples 0.0011564330000000001
choosing training 0.0011487860000000002
test sets 0.001132881
corresponding test 0.001132423
