word alignment 0.00395215
source word 0.0030102830000000003
word precision 0.0026694770000000004
single word 0.002632034
word align 0.002616634
content word 0.002607742
word lattice 0.002601557
word aer 0.00258975
identical word 0.002552649
arbitrary word 0.002533084
word sequences 0.002524688
word identity 0.002520243
word replacements 0.0025126610000000002
distinct word 0.002498426
tical word 0.0024936370000000004
unidirectional word 0.0024895240000000003
leverage word 0.0024895240000000003
word preci 0.0024895240000000003
word insertions 0.0024895240000000003
paraphrase sentence 0.00245229
target sentence 0.002375436
same sentence 0.002180493
source sentence 0.002174423
monolingual alignment 0.002145161
alignment pairs 0.002104842
alignment algorithm 0.002104286
alignment error 0.002090228
first sentence 0.002024707
phrasal translation 0.002023276
alignment algorithms 0.002007347
standard alignment 0.00200661
sequence alignment 0.002003219
automatic alignment 0.001980413
bidirectional alignment 0.001922706
large sentence 0.0018741810000000002
test sentence 0.0018698590000000002
sentence pair 0.001863608
sentence pairs 0.001849712
sentence set 0.001846572
machine translation 0.001786992
sentence length 0.001776971
language model 0.0017626529999999999
translation techniques 0.001748647
translation step 0.001742077
source words 0.001723258
input sentence 0.001722978
translation systems 0.001709183
training data 0.001700038
sentence rewrites 0.001689743
alignment 0.00168571
sentence types 0.0016759540000000001
sentence rank 0.001672185
data source 0.001661099
duplicate sentence 0.0016600640000000002
interesting sentence 0.0016572140000000002
alignable sentence 0.001655772
stereotypical sentence 0.001653546
translation quality 0.001630776
sophisticated translation 0.001626153
target language 0.0016198290000000001
translation database 0.001617581
phrase pair 0.001495668
paraphrase evaluation 0.0014909189999999998
phrase pairs 0.001481772
parallel data 0.0014653259999999999
model tion 0.0014394639999999999
ing paraphrases 0.001433781
sentence 0.00143058
same language 0.001424886
phrasal alignments 0.0014165010000000001
ing data 0.0014137120000000001
ibm model 0.001410289
possible paraphrases 0.001407093
paraphrase output 0.001402474
output paraphrase 0.001402474
msa paraphrase 0.001390541
translation 0.00138651
multiple data 0.0013857869999999999
replacement model 0.0013851
model scores 0.0013727169999999999
trigram model 0.00136562
baseline paraphrase 0.001360305
target sentences 0.001356953
training corpus 0.001351608
channel model 0.001335345
phrase research 0.001333628
phrase generation 0.0013318619999999998
additional data 0.001311672
msa paraphrases 0.0013061560000000002
news data 0.001306009
target length 0.001291247
paraphrase generation 0.0012909319999999998
lingual paraphrase 0.0012890979999999998
target languages 0.0012815
target side 0.001277579
paraphrase learning 0.0012758659999999998
paraphrase identification 0.0012514
paraphrase acquisition 0.0012511979999999998
paraphrase candidates 0.0012500769999999998
possible alignments 0.0012495030000000001
