translation model 0.0039296999999999995
alignment model 0.00349614
word alignment 0.00312149
word model 0.0031143100000000003
syntactic rules 0.003085498
target translation 0.003058858
syntactic translation 0.002995348
different rules 0.002914087
large rules 0.002879538
translation models 0.002865318
syntactic rule 0.002763998
translation system 0.002759628
possible rules 0.0026756230000000002
small rules 0.002666986
same rule 0.002651217
type rules 0.002630562
machine translation 0.0026067819999999998
motivated rules 0.002593507
different rule 0.002592587
weighted rules 0.0025676790000000002
large rule 0.002558038
translation experiments 0.002556908
useful rules 0.002554409
translation path 0.002550972
rules tune 0.002548429
order translation 0.002543488
extract rules 0.002536883
formed rules 0.0025320300000000002
translation task 0.002517445
direct translation 0.002493268
chine translation 0.0024452579999999996
translation pipeline 0.002442701
ern translation 0.0024416909999999997
translation etree 0.0024416909999999997
alignment models 0.002431758
rule set 0.0024263419999999997
rule probabilities 0.0024100939999999998
rule quality 0.002363096
possible rule 0.002354123
rule extraction 0.002338697
rule space 0.002332906
rule weight 0.0022969559999999997
rules 0.00227537
derivation tree 0.00227511
rule sequences 0.002228319
rule atoms 0.0022275999999999997
poor rule 0.002214316
english tree 0.0021965279999999997
first model 0.0021876820000000003
second model 0.002187265
translation 0.00218522
alignment link 0.002177745
alignment quality 0.002160886
syntax alignment 0.0021503119999999997
syntax model 0.002143132
ibm alignment 0.00212994
initial alignment 0.002124556
model space 0.002123516
ibm model 0.00212276
model ibm 0.00212276
correct alignment 0.002110759
bootstrap alignment 0.002110224
alignment procedure 0.002079264
partial alignment 0.002072573
motivated alignment 0.002069797
chinese word 0.002063099
tree decoder 0.00206234
weighted model 0.0020367890000000002
bad alignment 0.002031117
collins model 0.002024714
lation model 0.0020230120000000002
ment model 0.002013745
alignment qualities 0.002008452
possible tree 0.0019306129999999999
tion tree 0.0018402599999999998
tree spans 0.0018146549999999999
tree fragment 0.001799873
tree traversal 0.001787108
alignment 0.00175166
ibm word 0.00174811
model 0.00174448
null word 0.0016891270000000001
foreign word 0.001662653
viterbi word 0.001637146
chinese words 0.0015398669999999999
english words 0.001512766
training corpus 0.00146247
same data 0.001439787
training data 0.001403266
single probability 0.0013865610000000001
chinese bleu 0.001362895
large data 0.001346608
probability table 0.0013455070000000001
other features 0.0013380800000000002
probability prule 0.001304805
chinese string 0.001284333
new alignments 0.001281959
test data 0.001256872
combined probability 0.0012514639999999999
bootstrap alignments 0.001244466
