translation model 0.00511012
different translation 0.00460071
target phrase 0.00438436
source phrase 0.004268563
phrase table 0.004217205
phrase pairs 0.004213813
phrase combination 0.004170879
phrase reduction 0.004137011
translation rule 0.004086546
translation models 0.004086058
phrase number 0.004065061
phrase entries 0.004059189
same translation 0.004034312
translation score 0.004015146
translation results 0.003990578
similar translation 0.0039672990000000005
individual phrase 0.003921892
translation rules 0.0039169040000000006
translation performance 0.003909371
translation scores 0.003883903
translation evaluation 0.0038529700000000003
machine translation 0.003821112
translation coverage 0.0037159280000000003
translation knowledge 0.003705898
translation tasks 0.0037030070000000003
original translation 0.0036958990000000003
translation sys 0.0036778080000000003
translation perfor 0.003667773
unnecessary translation 0.003647384
phrase 0.0036153
translation 0.00338556
word alignment 0.0028309750000000003
training data 0.00265272
language model 0.0025466300000000002
single word 0.002481104
word align 0.0024536130000000003
automatic word 0.00244604
model size 0.002252806
smt model 0.002249135
model reduction 0.002246271
tion model 0.0021376249999999998
proposed model 0.001988878
baseline method 0.0017784659999999998
rate training 0.0017434170000000001
different reduction 0.001736861
model 0.00172456
source phrases 0.001718453
target words 0.001597457
target language 0.00159113
target sentence 0.001584596
reduction method 0.001573211
baseline system 0.0015406299999999999
bleu score 0.001493666
source sentence 0.0014687989999999998
training 0.00145888
other words 0.001450092
same target 0.001417812
sentence pairs 0.001414049
parallel sentence 0.001371255
bilingual sentence 0.001365427
bleu scores 0.0013624230000000002
dency method 0.001337965
dependency method 0.00133702
novel method 0.001335929
unigram phrases 0.001328896
cutoff method 0.0013172729999999999
method try 0.0013142009999999999
rule table 0.001302891
target sentences 0.0013002180000000001
feature function 0.0012897960000000002
rule reduction 0.001222697
parallel corpus 0.0012065140000000001
bilingual corpus 0.001200686
ibm models 0.001184042
hierarchical rule 0.001182488
original bleu 0.0011744190000000001
pruning entries 0.001113929
tion models 0.0011135630000000001
same time 0.001110724
development set 0.001109942
monolingual corpus 0.001077641
reduction task 0.001076029
trigram target 0.001073086
feature weights 0.001069318
other entries 0.001065584
phrases 0.00106519
rule redundancy 0.001060018
combination scores 0.001053922
method 0.0010515
size reduction 0.001049957
tion score 0.001042651
search space 0.0010401149999999999
opment set 0.001023798
hierarchical rules 0.001012846
performance evaluation 9.91221E-4
redundancy score 9.886180000000001E-4
last feature 9.77034E-4
rule num 9.71877E-4
erarchical rule 9.71436E-4
rule reduc 9.67635E-4
