translation model 0.0054182399999999995
translation models 0.004236403
translation rules 0.004064178
other translation 0.0039007110000000003
btg translation 0.003754214
reference translation 0.003737224
translation results 0.003715656
good translation 0.003705536
translation decoder 0.003658342
translation quality 0.0036235440000000002
current translation 0.003618414
translation result 0.003617585
bad translation 0.003615346
corresponding translation 0.003615027
translation candidates 0.00359735
final translation 0.003585786
translation ref 0.003568432
translation candi 0.00356791
translation 0.00329904
training data 0.002688821
model parameter 0.00253268
dsp model 0.002442721
bilingual training 0.002434696
guage model 0.002402985
source phrase 0.0023481360000000002
target phrase 0.002171294
word alignment 0.002120235
model 0.0021192
phrase pair 0.002056488
training instances 0.002050133
training process 0.002015052
detiled training 0.001993057
phrase length 0.0019227460000000001
different derivation 0.0019148340000000002
phrase pairs 0.001908019
decoding derivation 0.001897465
representation phrase 0.0018600560000000001
source sentence 0.001858033
word embedding 0.001837249
derivation space 0.001769612
training 0.00172486
noun phrase 0.001719044
used phrase 0.0017170010000000001
derivation structures 0.0015886490000000001
derivation set 0.0015836370000000002
derivation structure 0.001582333
source side 0.001490943
good derivation 0.001488456
test sentence 0.001487726
possible derivation 0.001461576
phrase 0.00144913
different derivations 0.001435295
smt decoding 0.001422544
entire derivation 0.001418135
decoding derivations 0.001417926
sentence pairs 0.001417916
gold derivation 0.001415442
chinese words 0.0014120699999999999
novel derivation 0.00140931
bad derivation 0.001398266
source phrases 0.001388235
input sentence 0.001386488
risk function 0.0013809640000000001
node vector 0.001374133
indicator function 0.001362672
derivation struc 0.001354802
complete derivation 0.0013533730000000002
chinese sentence 0.001352717
btg system 0.001352053
input source 0.001326467
target side 0.0013141009999999998
monolingual data 0.001295858
baseline system 0.001291146
source span 0.001290549
reordering limit 0.001289349
news data 0.001273858
local score 0.001267097
neural network 0.001263836
source spans 0.001230148
source string 0.001227102
different ones 0.001226522
target phrases 0.0012113929999999998
bilingual phrases 0.001199065
score figure 0.001185913
different gold 0.0011663560000000001
high score 0.0011607800000000001
node network 0.001154271
different ways 0.001148345
ing derivations 0.001124517
dsp score 0.001111946
different unn 0.0011023180000000001
force decoding 0.001101612
neural net 0.001099589
forced decoding 0.001097257
systems bleu 0.001089526
cal score 0.001084739
derivation 0.00108196
function 0.00107893
low score 0.001077954
final score 0.001075171
