translation output 0.0061743530000000005
other translation 0.006145797000000001
machine translation 0.006097565
english translation 0.005979985
translation consensus 0.0059616750000000005
translation candidate 0.005950408
candidate translation 0.005950408
translation memory 0.005947127
similar translation 0.005947018
translation performance 0.00594628
translation hypotheses 0.005941643
correct translation 0.00589152
translation pair 0.0058791320000000005
good translation 0.005868777
translation task 0.005861384
final translation 0.005834932
translation candidates 0.005819681
translation examples 0.005794694
translation posterior 0.005763581
leverage translation 0.005755368
translation 0.00546763
source sentence 0.0022951399999999998
training sentence 0.001861174
source sentences 0.001848394
similar source 0.001845408
training data 0.0018210029999999999
source span 0.0018044419999999999
source string 0.0017794339999999999
system 0.00168776
source spans 0.001667273
source strings 0.0016663259999999999
graph model 0.001618496
bleu score 0.001590401
bilingual training 0.0014776799999999999
first sentence 0.001473687
new rule 0.0014726510000000002
similar translations 0.001440064
other systems 0.0014298520000000001
sentence pairs 0.001383074
nist data 0.001377028
data set 0.0013731400000000001
rule updates 0.001370083
source 0.00136602
input sentence 0.001360936
data sets 0.001348662
ibm bleu 0.001339556
second sentence 0.001335798
sentence level 0.001316407
level bleu 0.001311716
complete sentence 0.001295497
training instances 0.001293959
training instance 0.001282275
multiple systems 0.001254415
model 0.00125028
translations ݁ᇱs 0.00125015
reference sentence 0.001242697
symmetrical sentence 0.00121714
decoder features 0.001168175
search space 0.001099804
final output 0.0010740250000000002
different instances 0.001068109
propagation probability 0.001047191
decoding process 0.001040448
other candidates 0.001030218
consensus features 0.0010252220000000001
conditional probability 0.0010006659999999999
second decoding 9.98141E-4
probability ܶሺሺ݂ 9.9388E-4
simple label 9.84039E-4
large graph 9.78999E-4
candidate selection 9.76166E-4
propagation algorithm 9.671840000000001E-4
similar sentences 9.61762E-4
translations 9.60676E-4
feature vector 9.50664E-4
training 9.32054E-4
learning method 9.303860000000001E-4
sentence 9.2912E-4
feature weights 9.26251E-4
new features 9.23438E-4
large dataset 9.03381E-4
structured learning 8.99653E-4
useful features 8.97637E-4
mbr decoding 8.953399999999999E-4
hypothesis mixture 8.944739999999999E-4
correct label 8.87817E-4
mixture decoding 8.855009999999999E-4
general case 8.845190000000001E-4
good candidate 8.83925E-4
collaborative decoding 8.81819E-4
structured label 8.809239999999999E-4
local consensus 8.808760000000001E-4
new hypotheses 8.66274E-4
loss function 8.547159999999999E-4
possible labels 8.51786E-4
simple category 8.432559999999999E-4
previous work 8.4317E-4
similar instances 8.41293E-4
experimental results 8.40823E-4
first author 8.38112E-4
