translation model 0.00554346
language translation 0.00526493
translation rule 0.00523252
translation system 0.005156646
other translation 0.004820301
possible translation 0.004719557
many translation 0.004719013
translation models 0.004665982
translation systems 0.004644755
maximum translation 0.004618378
string translation 0.004574999
translation quality 0.004567835
translation train 0.004553748
translation stage 0.004547375
translation sacrifices 0.00450515
translation metric 0.004502407
translation derivations 0.004486238
translation deriva 0.0044787030000000005
translation metrics 0.004476431
translation yields 0.004476431
translation 0.0041992
source source 0.0025194
language model 0.00240999
source language 0.00232543
source tree 0.002235084
same model 0.0020245389999999997
source target 0.001997338
same source 0.001939979
word alignments 0.001928868
training data 0.001869517
model size 0.001861876
source side 0.001848755
syntactic parser 0.0018233070000000001
lation model 0.0018091189999999999
parallel data 0.0017507120000000002
same rules 0.0016547340000000002
data size 0.001637226
source trees 0.0016343050000000002
trigram model 0.001624258
syntactic structure 0.001617219
unique source 0.001590517
side tree 0.001564439
parse tree 0.001556892
source lan 0.001554976
language text 0.001482821
parser probability 0.0014760770000000001
maximum rule 0.001452498
language string 0.001441529
system score 0.001428993
single tree 0.001383811
rule extraction 0.0013721889999999998
structure parser 0.0013609260000000001
correct tree 0.001358396
rule height 0.001353983
string tree 0.0013511830000000002
tree string 0.0013511830000000002
model 0.00134426
string system 0.001333245
syntactic unit 0.00132718
baseline parser 0.001313846
same features 0.001297809
other words 0.001281211
source 0.0012597
system objective 0.001255437
bleu score 0.0011966
correct parser 0.001166519
new feature 0.001155238
original parser 0.001152599
chinese parser 0.0011367249999999999
parser stage 0.001131682
other hypothesis 0.00112489
results table 0.0011190599999999998
ing performance 0.001114804
target string 0.001113437
parse forest 0.001103648
possible parse 0.001101865
other information 0.001093988
parser errors 0.0010925940000000001
mer training 0.001072902
standard algorithm 0.001070657
charniak parser 0.001065977
language 0.00106573
parser gen 0.0010639920000000001
nese parser 0.001062721
string figure 0.001057466
parallel text 0.001048193
main results 0.001046508
bilingual features 0.001045935
head words 0.0010325339999999999
previous work 0.001004088
single parse 9.89935E-4
feature weights 9.82168E-4
posterior probability 9.7552E-4
rules 9.74455E-4
ing accuracy 9.59122E-4
system 9.57446E-4
parse trees 9.56113E-4
recent work 9.53971E-4
side structures 9.46548E-4
chinese side 9.42273E-4
