translation model 0.0038528300000000002
language model 0.002764274
translation performance 0.0026664640000000003
translation rules 0.002643016
reordering model 0.002564848
translation models 0.002517531
phrase translation 0.0024971100000000003
reference translation 0.002482353
reordering features 0.002478228
training data 0.00243389
syntactic features 0.002423219
probability features 0.002404863
different features 0.002402639
model score 0.002392046
constituency features 0.002327744
machine translation 0.0023254960000000002
hypothesis translation 0.0023195810000000002
translation task 0.002308915
feature set 0.002297698
translation probabilities 0.002286715
translation quality 0.002254064
oracle translation 0.002251548
new features 0.002233921
baseline model 0.002232689
translation hypotheses 0.0022286320000000004
rate model 0.0022200790000000002
hierarchical model 0.002214598
first features 0.002211885
cle translation 0.0022117770000000003
constituency feature 0.002204104
linear model 0.00219967
distortion model 0.002176577
ibm model 0.0021658330000000003
binary feature 0.0021536710000000002
feature weights 0.002150717
baseline features 0.002146069
aware model 0.0021140160000000002
new feature 0.002110281
individual features 0.002098853
syntax features 0.002094548
distortion features 0.002089957
hiero features 0.002070887
various feature 0.002068019
feature sets 0.002048824
separate features 0.002038818
sensitive features 0.002037996
tortion features 0.002027896
baseline feature 0.002022429
single feature 0.00202199
second feature 0.002012776
feature combinations 0.001977215
translation 0.0019753
feature definition 0.001946524
training algorithm 0.001934224
feature subset 0.0019285700000000001
training sentence 0.001923264
separate feature 0.001915178
model 0.00187753
test data 0.001868093
tuning data 0.0018442379999999998
features 0.00179091
training method 0.0017880370000000001
evaluation data 0.001709022
source language 0.0017041069999999998
feature 0.00166727
development data 0.0016368049999999999
other word 0.001622048
discriminative training 0.0015979380000000001
source words 0.0015524129999999999
treebank data 0.0015390199999999999
rate training 0.001535769
word alignment 0.001511156
training methods 0.0015089460000000002
uation data 0.001478115
language models 0.0014289749999999999
loss function 0.001413044
first word 0.001388257
language pair 0.001387044
new language 0.001329755
language side 0.00127999
objective function 0.001277231
many source 0.0012737410000000001
language pairs 0.001267376
word penalty 0.001266266
test set 0.001257851
reference sentence 0.001237097
tuning set 0.001233996
such models 0.001220363
source side 0.001210609
phrase reordering 0.001209128
grammar rule 0.0012009
function clip 0.001197667
training 0.00119322
reference translations 0.001193094
different sentences 0.001182374
learning algorithm 0.001152523
system scores 0.0011067030000000001
single sentence 0.001084764
same set 0.0010740860000000001
new set 0.001073439
