sentence translation 0.00358318
translation system 0.003398727
language model 0.0033907
translation probabilities 0.003098358
machine translation 0.003089248
translation table 0.003085376
training data 0.0029493839999999998
lexical translation 0.002897925
twitter translation 0.002895749
translation options 0.002857648
translation probabil 0.002827813
parallel data 0.0028244539999999997
data set 0.002626635
language sentence 0.00259245
domain data 0.002590505
model score 0.002585496
translation 0.0025666
retrieval model 0.002550564
bilingual data 0.0024961159999999996
domain model 0.002450505
evaluation data 0.002405628
wikipedia data 0.0023864809999999998
new data 0.0023861729999999997
language sentences 0.002374998
clir model 0.002365977
language query 0.002357924
nist data 0.0023380469999999998
smt model 0.002312414
twitter data 0.002283979
gual data 0.0022622939999999998
tracted data 0.0022550499999999998
previous model 0.0022527099999999998
comparable data 0.002251366
parable data 0.002217738
model com 0.0021248589999999998
trieval model 0.0020940869999999997
target language 0.002080052
source language 0.002079551
model mgen 0.002075422
language tweets 0.002023676
parallel sentence 0.001886204
language modeling 0.0018826250000000002
language identification 0.001841811
language tweetsdtrg 0.001835407
language tweetsqsrc 0.001835407
model 0.00181483
sentence retrieval 0.001752314
parallel sentences 0.001668752
word alignments 0.0016131980000000001
language 0.00157587
english translations 0.001570142
retrieval approach 0.001526844
sentence pairs 0.001454267
query set 0.001453859
learning method 0.001446543
large corpus 0.001445741
baseline system 0.001374899
sentence extraction 0.001356791
clir approach 0.001342257
test tweets 0.001338512
additional training 0.0013099
trieved sentence 0.001278176
quality parallel 0.001243385
query term 0.001214882
full sentences 0.001212429
small set 0.001211584
parallel sen 0.001202435
parallel twitter 0.001198773
twitter corpus 0.001191921
retrieval results 0.001188822
unknown words 0.001183475
smt models 0.001177071
weighted translations 0.0011722360000000001
standard domain 0.001159985
comparable corpus 0.001159308
test split 0.0011503120000000001
english tweets 0.0011401599999999999
different ways 0.001139131
retrieval step 0.001138806
retrieval performance 0.001137908
standard features 0.0011261230000000001
corpus dtrg 0.001122479
different development 0.001113272
different decoder 0.001113059
target text 0.001112923
english documents 0.001102613
candidate sentences 0.0011016189999999999
ing smt 0.001097916
tion retrieval 0.001086706
different variations 0.0010850019999999998
broadcast news 0.001081617
allel sentences 0.001078624
transductive learning 0.001073143
learning curves 0.001072931
query terms 0.0010720550000000001
trieval approach 0.001070367
iterative approach 0.001067716
particular query 0.001062029
sentences sgen 0.001061241
general domain 0.0010608
