word alignment 0.007365450000000001
alignment model 0.005185840000000001
alignment links 0.00515156
alignment training 0.005004426
alignment features 0.004992502
alignment models 0.004873145000000001
alignment link 0.004828745000000001
alignment combination 0.004680184
alignment selection 0.004633508000000001
baseline alignment 0.004632932
word alignments 0.004609520000000001
alignment quality 0.004543767000000001
discriminative alignment 0.004533653
final alignment 0.004527435000000001
reference alignment 0.0044553050000000005
combined alignment 0.004444443
alignment combi 0.004436964
alignment outper 0.004428392000000001
alignment 0.00412995
word align 0.0039662180000000005
target word 0.0037211880000000003
source word 0.003668081
partial word 0.003661508
pashto word 0.00365109
neighboring word 0.0035666870000000002
complementary word 0.003565533
diversified word 0.003544817
translation probability 0.0023794719999999997
lexical translation 0.0023453249999999997
translation task 0.002297121
translation performance 0.002270104
english words 0.0022585
machine translation 0.002243167
pashto translation 0.00218014
english sentence 0.0020050479999999997
additional alignments 0.0018494520000000001
training data 0.0018054
translation 0.00176455
sentence set 0.001706702
english sentences 0.0016953659999999998
combined alignments 0.001688513
diversified alignments 0.001683337
individual alignments 0.001673741
english parse 0.001592683
partial words 0.001558448
phrase training 0.0015561820000000001
partial english 0.001552068
pashto words 0.00154803
same training 0.001542421
different align 0.001500441
test set 0.0014988240000000002
english side 0.0014986659999999998
entropy model 0.001482564
candidate links 0.0014677870000000001
different approach 0.00146044
english constituent 0.0014340029999999999
urdu words 0.001432675
search method 0.001430553
single set 0.001426006
sentence pairs 0.001393022
link posterior 0.001386507
alignments 0.00137402
block model 0.00135544
posterior probability 0.001302634
sentence pair 0.001301969
pashto sentence 0.001294578
multiple align 0.001263793
data sparsity 0.001262925
rate training 0.001253416
our training 0.001250336
combination approach 0.0012409510000000001
baseline align 0.0012337
glish sentence 0.001198663
selection approach 0.0011942749999999999
bleu scores 0.0011825870000000001
fixed set 0.001176929
conversational training 0.001174098
different sets 0.001158769
tuning set 0.001158697
candidate link 0.001144972
words 0.00113244
english 0.00112606
different directions 0.001122621
different aligners 0.001114921
simple method 0.001107162
sov languages 0.001091361
decoding weights 0.001083803
novel method 0.001073338
different criterion 0.001070248
different motivations 0.001070248
model 0.00105589
combination algorithm 0.001055649
syntactic rules 0.001034594
linguistic knowledge 0.001023226
statistical machine 0.00102206
links 0.00102161
other sets 0.001021126
link ajk 9.99576E-4
greedy search 9.92733E-4
posterior probabil 9.88418E-4
