training training 0.00393062
translation phrase 0.0037771000000000002
translation model 0.0034709
bleu training 0.00322214
translation models 0.002988992
training data 0.002947186
translation probability 0.0028952120000000003
baseline translation 0.002801202
translation system 0.002742975
machine translation 0.0026611440000000003
network training 0.002647899
linear translation 0.0026035200000000002
ing translation 0.0025581180000000003
same translation 0.0025355290000000003
translation tasks 0.002468032
reference translation 0.002451721
translation sys 0.002435039
particular translation 0.002374769
chine translation 0.0023683610000000003
rate training 0.0023429790000000002
source words 0.00233862
training example 0.002328316
training times 0.002301931
training complex 0.002292063
language model 0.002211625
translation 0.00209608
network model 0.002057409
model score 0.001985944
training 0.00196531
english word 0.0019455240000000001
next word 0.001910094
same word 0.001899399
bleu loss 0.0018783480000000002
test data 0.001870676
bleu score 0.001867954
word scores 0.0018375800000000001
ing model 0.0018368579999999998
previous words 0.00180962
input words 0.001765497
individual word 0.001739596
language models 0.001729717
bleu objective 0.001729511
trained model 0.001709417
phrase 0.00168102
next words 0.001662554
guage model 0.001660964
expected bleu 0.001641989
neural network 0.00163087
average bleu 0.001614419
bleu variant 0.001593292
error function 0.001592946
data set 0.00158136
network models 0.001575501
same source 0.001565659
bleu approximation 0.001564448
level bleu 0.001564125
bleu mod 0.001556796
loss function 0.0015531960000000002
network language 0.001519394
output layer 0.001515321
recurrent neural 0.001513573
test time 0.00145911
recurrent models 0.001458204
same data 0.001421325
input layer 0.001409907
source span 0.001399686
parallel data 0.001380752
model 0.00137482
language pair 0.001370231
hidden layer 0.001355584
baseline system 0.001352017
layer size 0.001343259
data sets 0.001309162
network score 0.0012937130000000002
neural networks 0.001287268
neural net 0.0012732820000000001
component models 0.001273058
allel data 0.001256802
lel data 0.0012505020000000001
recurrent network 0.0012478810000000002
tion models 0.001240655
language pairs 0.00123885
activation function 0.00123807
complex models 0.001219665
test sets 0.001216086
words 0.00121241
common features 0.001209282
previous work 0.001207813
domain test 0.0011978190000000001
time algorithm 0.0011946629999999999
previous time 0.00116752
single sentence 0.001146851
layer state 0.001144509
distortion feature 0.0011412829999999999
layer configuration 0.00114047
language mod 0.001136771
put layer 0.0011338680000000001
layer configurations 0.001129929
den layer 0.001127986
source 0.00112621
