paraphrase sentence 0.0033709300000000003
paraphrase corpus 0.00320528
different paraphrase 0.0030995930000000003
translation model 0.0029959569999999996
paraphrase evaluation 0.002876205
machine translation 0.002851794
translation system 0.002827637
translation results 0.002784488
paraphrase results 0.002731218
translation systems 0.002730329
paraphrase generation 0.002665619
monolingual translation 0.002661428
paraphrase quality 0.002648708
translation process 0.0026171
candidate paraphrase 0.002563435
paraphrase adequacy 0.002517315
paraphrase rate 0.002507582
optimal paraphrase 0.002499414
translation engines 0.002491013
translation candidates 0.002490259
translation tables 0.002485111
translation processes 0.0024848509999999997
paraphrase purposes 0.002482224
paraphrase result 0.002457214
paraphrase criteria 0.002441471
paraphrase ade 0.002434097
paraphrase gener 0.002433418
paraphrase qual 0.002430529
paraphrase genera 0.00242772
translation 0.00223342
paraphrase 0.00218015
language method 0.001924524
english sentence 0.001914504
language model 0.001817117
evaluation data 0.0018125250000000002
source sentence 0.001809352
human evaluation 0.001791755
same language 0.001779952
sentence pairs 0.001741753
chinese sentence 0.001735211
input sentence 0.001729836
parallel data 0.001702654
human translations 0.001692959
sentence quality 0.001659338
sentence level 0.001630447
original sentence 0.0016133599999999999
parallel corpus 0.001611314
good sentence 0.001592103
pivot language 0.001571219
learning method 0.001566635
data set 0.001560249
ing data 0.001542002
language generation 0.001540049
phrase results 0.001533612
smt model 0.001523445
sentence pair 0.001508292
tion data 0.0014756510000000001
phrase quality 0.001451102
glish sentence 0.001443752
sentence levels 0.00143563
development data 0.001412929
bleu scores 0.001405253
different words 0.001404134
human evaluations 0.001401619
human judgments 0.001375792
phrase candidate 0.0013658289999999998
testing data 0.00136308
smt system 0.001355125
source sentences 0.001353817
human eval 0.00134863
natural language 0.0013293900000000002
learning bleu 0.001318287
different application 0.001301854
inal language 0.0012993890000000002
bleu score 0.0012788399999999998
smt systems 0.001257817
evaluation results 0.0012471230000000002
phrase tables 0.001234235
automatic evaluation 0.00121619
different settings 0.001211163
score results 0.001208312
smt features 0.001207793
different applications 0.0011995880000000001
same meaning 0.001196562
sentence 0.00119078
different applica 0.00116792
different priorities 0.001164944
different demands 0.001164944
different pivots 0.001164944
inference method 0.001162104
original sentences 0.001157825
evaluation metric 0.001120635
joint learning 0.0011154960000000001
same time 0.001095646
learning process 0.0010803710000000001
short sentences 0.0010731310000000002
tion evaluation 0.001055236
language 0.00105458
dual smt 0.00105017
bleu metric 0.001046176
