translation model 0.002566409
translation process 0.002170514
english translation 0.002150299
translation probability 0.002043257
translation problem 0.002040865
standard translation 0.001996948
results translation 0.001950561
translation accuracy 0.001928547
tree fragment 0.0019194260000000001
candidate translation 0.001899658
translation quality 0.001865302
optimal translation 0.001842636
sfr translation 0.001835852
translation output 0.00183386
line translation 0.001818673
translation decisions 0.001815075
poor translation 0.00180913
correct translation 0.00180708
devset translation 0.001806392
mize translation 0.001806392
language model 0.001730657
same fragment 0.0017065280000000001
fragment trees 0.00163794
translation 0.00158355
tree fragments 0.001558092
fragment score 0.001550911
source fragment 0.001540802
target fragment 0.001519089
fragment scores 0.001510318
dot fragment 0.001475023
other nodes 0.001473565
source grammar 0.001471692
fragment pair 0.001467009
fragment order 0.001454094
other fragments 0.001427841
new rule 0.001424578
fragment accuracy 0.001423537
fragment pairs 0.001419867
dot grammar 0.001405913
candidate fragment 0.001394648
particular fragment 0.001391035
parent fragment 0.001387803
dot model 0.001379342
fragment rescoring 0.001372537
fragment edit 0.001370581
baseline grammar 0.0013654679999999999
fragment quality 0.001360292
oracle fragment 0.001359696
language sentence 0.0013568550000000001
fragment surface 0.001353458
fragment yields 0.001350995
fragment difference 0.001342362
structured fragment 0.001324474
rule scores 0.001318976
unary fragment 0.001317029
fragment aligner 0.001304698
fragment scoring 0.001303849
source tree 0.001303148
fragment roots 0.001302865
fragment dgs 0.001301363
linked fragment 0.001301363
grammar size 0.0012882599999999998
training sentence 0.00126922
parallel tree 0.0012660409999999999
grammar output 0.00125974
channel model 0.0012419129999999999
rescored grammar 0.00124063
lation model 0.001230891
tree pair 0.0012293550000000001
same set 0.0012147310000000001
syntactic features 0.00118888
target language 0.001188347
tree pairs 0.001182213
other improvement 0.001162487
minimum tree 0.001161155
root node 0.001160117
target fragments 0.001157755
new sentence 0.001146437
training corpus 0.001145404
scores translations 0.001143731
feature weights 0.001130977
ing translations 0.001129489
frontier node 0.00111941
pcfg rules 0.001119333
node pairs 0.001117105
rule instance 0.0011149439999999999
dot fragments 0.001113689
frontier nodes 0.001106562
additional language 0.0011018310000000002
small fragments 0.001090696
same source 0.00109025
search problem 0.001087964
training data 0.00107952
fragment 0.00107854
source sentence 0.001071319
tree transformations 0.001063561
internal nodes 0.0010621860000000001
word order 0.001061668
individual rules 0.001048827
training method 0.00104162
