paraphrase phrase 0.00326818
paraphrase data 0.00284315
language phrase 0.002651319
target translation 0.002601
paraphrase model 0.00255906
translation process 0.0024967550000000002
chinese translation 0.002476318
phrase table 0.002460438
source phrase 0.0024553459999999997
correct translation 0.0024508430000000003
translation quality 0.002447756
phrase alignment 0.002437655
translation result 0.002416653
english phrase 0.002388425
flexible translation 0.002387007
paraphrase test 0.002383276
translation tasks 0.002380248
translation difficulty 0.002379207
translation options 0.0023746550000000003
translation difficult 0.0023724880000000003
translation diffi 0.002369838
exact translation 0.0023693760000000003
translation unit 0.0023687490000000003
fundamental translation 0.0023687490000000003
phrase pairs 0.002301833
training data 0.0022762859999999998
phrase extraction 0.002249286
paraphrase words 0.002226285
phrase probabilities 0.002214128
phrase substitution 0.002191596
translation 0.00214034
phrase length 0.002129764
side phrase 0.002112907
maximum phrase 0.002099396
paraphrase table 0.002071458
phrase tables 0.002067037
known phrase 0.00206503
lar phrase 0.002064703
latter phrase 0.002060135
inaccurate phrase 0.002056677
news data 0.002051954
source word 0.002051946
word alignment 0.002034255
parallel data 0.0019965
model training 0.001992196
original word 0.0019203290000000001
data set 0.0019105939999999998
paraphrase lattice 0.0018901900000000001
test sentence 0.001887111
word lattice 0.0018757700000000001
phrase 0.00182858
paraphrase substitution 0.0018026160000000002
training corpus 0.0017773490000000001
paraphrase edges 0.00176551
final word 0.0017483070000000002
ing data 0.0017449009999999999
data sets 0.001731167
candidate paraphrase 0.0017285870000000002
data increases 0.001705822
word lattices 0.0016986430000000001
fbis data 0.001697973
word lat 0.001692177
limited data 0.001684477
paraphrase options 0.001673915
word sequences 0.0016724740000000002
paraphrase acquisition 0.0016704760000000002
paraphrase penalty 0.001668758
data conditions 0.001667925
data sparseness 0.001664587
allel data 0.001632643
data preparation 0.0016315359999999998
data availability 0.0016315359999999998
other language 0.001579685
potential paraphrases 0.001579417
source sentence 0.0015702009999999998
possible paraphrases 0.001559822
new model 0.001541738
ing paraphrases 0.001527461
parallel corpus 0.001497563
parallel training 0.001465686
test set 0.00145072
tactic paraphrases 0.001442187
few paraphrases 0.001442177
corresponding paraphrases 0.001440884
paraphrase 0.0014396
original sentence 0.001438584
acquisition paraphrases 0.001416986
other news 0.00140535
language models 0.00136887
input sentence 0.001367135
language translations 0.00135833
inal model 0.001350284
source phrases 0.001346866
language pairs 0.001295992
new training 0.0012950140000000001
target language 0.001283399
original words 0.001281834
english phrases 0.001279945
paraphrased sentence 0.001270561
training sen 0.0012419290000000001
