translation model 0.0033763099999999996
mbr translation 0.0029413679999999998
syntactic translation 0.0029217899999999996
translation forest 0.0028093149999999997
translation performance 0.0025699309999999997
machine translation 0.0025563689999999997
hierarchical translation 0.002554597
translation output 0.002540921
reference translation 0.002538579
translation system 0.002529987
consensus translation 0.002525926
baseline translation 0.00250381
viterbi translation 0.0024899479999999996
translation systems 0.002483448
forests translation 0.002448356
translation forests 0.002448356
candidate translation 0.0024453979999999997
risk translation 0.0024415129999999998
translation quality 0.0024396419999999997
translation tasks 0.002412255
correct translation 0.002404008
translation distribu 0.0024024569999999998
translation 0.0021582
model features 0.002106912
language model 0.001998821
linear model 0.001797281
mbr decoding 0.001713796
sentence features 0.001706127
model weights 0.0016662449999999998
syntactic features 0.001652392
time algorithm 0.0016355330000000002
statistical model 0.0016206859999999999
model score 0.0016166929999999998
forest decoding 0.001581743
model scores 0.0015658929999999998
mbr translations 0.001524421
total model 0.0015030339999999999
guage model 0.001478956
forest bleu 0.001472525
new algorithm 0.001464671
est model 0.0014635809999999998
core model 0.001462368
many features 0.001417601
other language 0.001391531
same decoding 0.001389918
linear mbr 0.001362339
decoding process 0.001357239
decoding performance 0.001342359
optimal decoding 0.001328018
sentence similarity 0.001317087
decoding procedure 0.001312783
decoding algorithms 0.001309571
consensus decoding 0.001298354
decoding table 0.0012941749999999998
bleu improvements 0.001289025
data set 0.001277098
fast algorithm 0.00127531
standard bleu 0.0012691949999999999
target sentence 0.0012687520000000002
possible translations 0.00126526
training data 0.001256948
small bleu 0.001253112
corpus bleu 0.0012519369999999998
word sub 0.001248714
word lattices 0.001246969
arabic sentence 0.0012417910000000001
eral algorithm 0.001234635
language models 0.001234132
maximum sentence 0.001230957
forest linear 0.001230286
list mbr 0.001226429
input sentence 0.001220241
model 0.00121811
risk decoding 0.001213941
feature similarity 0.001207356
decoding objective 0.001205166
indicator features 0.0012030769999999999
sentence pairs 0.001201768
sentence length 0.0011992040000000002
decoding framework 0.001184235
mbr speed 0.001180288
decoding procedures 0.0011768339999999999
syntactic parse 0.0011760709999999999
sensus decoding 0.001174726
bleu scores 0.001169193
language pairs 0.001165154
frequency features 0.0011545589999999999
fractional features 0.001149433
relevant features 0.001145586
dependency language 0.0011317950000000001
feature space 0.00113123
output translations 0.001123974
reference translations 0.001121632
ing feature 0.001121227
mbr approach 0.001108037
forest pruning 0.001103183
final translations 0.001102373
linear function 0.001099809
third language 0.001094891
explicit syntactic 0.0010881
