phrase translation 0.00771828
target phrase 0.005602472
phrase pair 0.005534647
english phrase 0.0054029080000000005
phrase length 0.005390307
models phrase 0.005381502
phrase table 0.005374744
pivot phrase 0.005359758
phrase pivot 0.005359758
arabic phrase 0.005311044
source phrase 0.005236665
phrase pairs 0.00515177
phrase probabilities 0.005139479000000001
new phrase 0.005088985
corpora phrase 0.004982273000000001
maximum phrase 0.00496863
phrase tables 0.004947587
phrase pivoting 0.004926204
equivalent phrase 0.004923958
sian phrase 0.00492301
alent phrase 0.00492301
translation model 0.0048851
phrase 0.00463618
sentence translation 0.00400566
translation models 0.003827422
table translation 0.003820664
pivot translation 0.003805678
arabic translation 0.003756964
translation probabilities 0.003585399
translation technique 0.003481858
translation quality 0.0034578309999999998
word alignment 0.00345721
translation mod 0.003437845
good translation 0.003427042
overall translation 0.003422532
translation prob 0.003421575
translation candidates 0.003383656
chine translation 0.003372356
translation choices 0.003369375
arabic word 0.003087974
translation 0.0030821
word corpus 0.003053608
language model 0.00285637
persian word 0.002766332
model bleu 0.002569725
training data 0.00239991
target phrases 0.002277902
model weights 0.002260155
model size 0.0022237719999999997
english data 0.002059058
different language 0.002044488
pivot phrases 0.002035188
target language 0.002019662
parallel data 0.0019974019999999997
language sentence 0.00197693
language pair 0.001951837
target words 0.0018262600000000001
model 0.001803
language models 0.0017986920000000002
pivot language 0.001776948
arabic language 0.001728234
candidate phrases 0.001706803
target languages 0.001643273
arabic target 0.001641156
data scarcity 0.001608338
target corpus 0.0016067899999999999
many language 0.001587888
allel data 0.001581865
language pairs 0.00156896
arabic words 0.0015348319999999999
source sentence 0.001524045
reordering models 0.001502429
reordering table 0.001495671
language modeling 0.0014944300000000002
source words 0.001460453
smt system 0.0014121329999999999
autonomous language 0.001387285
third language 0.001360756
bridge language 0.001360319
alignment links 0.00135482
smt models 0.001353787
parallel corpus 0.00134557
glish language 0.0013425580000000002
language families 0.0013421330000000001
language exploitation 0.001340585
phrases 0.00131161
target connectivity 0.001301846
different ways 0.001285534
arabic smt 0.001283329
persian sentence 0.001276782
equivalent target 0.00125407
same set 0.0012451110000000001
case system 0.001240306
meteor sentence 0.0012224100000000002
sentence pivoting 0.001213584
persian words 0.00121319
other features 0.001210781
first example 0.0011892299999999999
test set 0.0011871630000000002
bad pair 0.0011858300000000001
