translation model 0.0036510199999999996
alignment model 0.0034148399999999997
edit model 0.003352626
sequence model 0.003010306
smt model 0.002921504
generative model 0.002904503
reordering model 0.002840492
lexicon model 0.002830687
model parameters 0.002809667
order model 0.002777293
transcription model 0.002776775
model task 0.002752634
spelling model 0.00271242
current model 0.002694577
model precision 0.00266586
ranking model 0.0026434939999999997
ibm model 0.002641399
model parameterization 0.002611132
model simi 0.002603957
reliable model 0.002603158
model 0.00238304
source word 0.002128249
phrase translation 0.001942339
alignment models 0.001908488
training set 0.00188894
translation system 0.0018082049999999998
rhyming word 0.001723506
training algorithm 0.001683
machine translation 0.0016671309999999999
sequence alignment 0.001659066
simple features 0.001656476
target language 0.00162267
edit sequence 0.0015968520000000002
other target 0.001592151
perceptron training 0.001570696
target sequence 0.001560756
phrase models 0.001551047
other words 0.001535793
target string 0.0015351380000000001
training rule 0.001520577
training procedure 0.001515345
chine translation 0.001492148
target character 0.001477955
generation training 0.001470516
training regime 0.001469706
data set 0.001463923
fold training 0.00145065
training modes 0.00145065
standard edit 0.001440576
edit distance 0.001409466
generative models 0.001398151
sequence alignments 0.0013979180000000002
arbitrary features 0.0013928340000000001
indicator features 0.0013838000000000001
english words 0.00137185
possible target 0.0013715720000000002
preceding features 0.0013684090000000001
edit weights 0.00136791
gram features 0.001366716
arabic words 0.001343106
correct alignment 0.001342518
same set 0.0013286489999999999
source language 0.0013181690000000001
tion edit 0.0013065940000000001
target strings 0.0013024450000000002
latent alignment 0.001296905
alignment solution 0.001288708
particular edit 0.001282169
translation 0.00126798
feature values 0.001262778
quence alignment 0.001262654
edit cost 0.001261768
feature vector 0.001261461
edit affinity 0.001259015
corresponding edit 0.001258851
optimal edit 0.001258166
target sequences 0.001256343
source sequence 0.001256255
explicit alignment 0.001251794
alignment affinities 0.001250931
timal alignment 0.00124979
corresponding feature 0.001246869
several models 0.0012388570000000001
language character 0.001233645
training 0.00123244
source string 0.0012306370000000001
known edit 0.00121149
feature selection 0.001204332
feature vectors 0.001203761
transcription data 0.001201158
true target 0.001199949
reference target 0.0011966190000000001
edit costs 0.001193251
edit affinities 0.0011887170000000002
edit decisions 0.001188146
primitive edit 0.001187343
feature construction 0.001182254
target selection 0.001180218
target alphabet 0.001175513
source character 0.001173454
