sentence paraphrase 0.0046245
paraphrase pairs 0.003906277
same paraphrase 0.003891238
paraphrase lattice 0.003797071
translation phrase 0.0036033000000000003
phrase translation 0.0036033000000000003
paraphrase length 0.0035996739999999998
paraphrase probability 0.003572377
paraphrase list 0.0035701139999999997
paraphrase pair 0.003491371
proper paraphrase 0.003490452
paraphrase lat 0.003473408
paraphrase lattices 0.003472327
building paraphrase 0.003453594
erroneous paraphrase 0.0034442979999999997
neous paraphrase 0.0034393239999999997
paraphrase 0.0031382
phrase pairs 0.0026467269999999998
same phrase 0.002631688
phrase table 0.002443652
original phrase 0.0024436279999999998
phrase list 0.0023105639999999998
machine translation 0.0022626400000000002
appropriate phrase 0.002249277
translation experiments 0.002232499
source sentence 0.002230474
phrase lattices 0.002212777
language model 0.00220354
priate phrase 0.002180145
augmented phrase 0.002180145
hierarchical phrase 0.002180145
translation probability 0.0021588270000000003
translation quality 0.0021438710000000003
chinese translation 0.00212035
speech translation 0.002105654
input sentence 0.002100727
smt model 0.00208828
original sentence 0.002051278
turkish translation 0.0020267980000000002
parallel corpus 0.001984678
sentence length 0.0019477740000000002
paraphrased sentence 0.001926671
model score 0.0019081839999999998
proper paraphrases 0.001904172
phrase 0.00187865
output sentence 0.001840868
different phrases 0.00181566
source smt 0.001812294
other smt 0.001777556
smt system 0.001729665
translation 0.00172465
training corpus 0.0017098859999999999
web corpus 0.001680183
german word 0.001673767
word segmentation 0.001662561
training data 0.001661626
source sentences 0.001652665
compound word 0.0016108849999999998
bleu score 0.001598561
language models 0.0015548749999999998
paraphrases 0.00155192
europarl corpus 0.0015442339999999998
corpus size 0.001528146
input sentences 0.0015229179999999998
language texts 0.001516656
natural language 0.00151165
normalized language 0.001500302
list smt 0.0015000340000000001
sentence 0.0014863
smt techniques 0.0014650940000000001
standard smt 0.0014649650000000001
model probability 0.001454337
allel corpus 0.0014338239999999998
massive corpus 0.0014314689999999999
smt overview 0.0014062340000000001
smt features 0.001401059
work lattice 0.001392426
model proba 0.001369155
guage model 0.001338805
novel method 0.001335152
sentences sets 0.001332472
lattice decoding 0.0012965770000000001
unknown phrases 0.001296191
paraphrased phrases 0.001273932
same meaning 0.001242164
proposed method 0.001234827
experimental results 0.001214596
language 0.00118338
source sen 0.001182485
rate training 0.0011636799999999998
decoding figure 0.00115416
bleu scores 0.001137211
corpus 0.00113004
same way 0.001107329
other nodes 0.001097291
other hand 0.001089017
future work 0.001077689
decoding feature 0.00107485
related work 0.001068502
open source 0.001066384
