translation model 0.002779601
possible translation 0.002499174
translation quality 0.002324618
good translation 0.002301133
translation systems 0.002293219
machine translation 0.0022497710000000002
translation accuracy 0.002239537
chinese translation 0.002189416
cube pruning 0.0021643820000000003
pruning cube 0.0021643820000000003
beam search 0.002162448
translation 0.00190068
following rule 0.001897843
decoding time 0.001827785
search error 0.001808781
search quality 0.001771848
scfg rule 0.001714503
same source 0.00170945
search space 0.0017029529999999999
potential search 0.001694382
model cost 0.00169266
search errors 0.0016909289999999999
various pruning 0.0016486510000000001
source word 0.001636462
pruning settings 0.001625937
search effi 0.0016165379999999998
search spaces 0.0016165379999999998
pruning alg 0.0016142090000000001
different model 0.0015929999999999998
node cube 0.00158275
language model 0.0015570699999999998
multiple rules 0.001505532
parsing algorithm 0.0014691230000000001
different target 0.001423498
same test 0.0013906629999999999
source side 0.001379687
search 0.00134791
pruning 0.00134577
possible items 0.0013383129999999998
different models 0.001322455
average model 0.001315452
efficient decoding 0.0013141379999999999
source string 0.001300022
above algorithm 0.001287173
language models 0.001286525
same level 0.001279096
data set 0.001272155
antecedent nodes 0.001270324
same coverage 0.001260033
target side 0.001252315
model costs 0.001246408
node table 0.001238791
future item 0.001217709
tecedent nodes 0.001211651
last phrase 0.001209933
visits nodes 0.001209085
combination cost 0.001203161
trigram model 0.001185728
gram model 0.001182339
target string 0.00117265
true cost 0.0011699319999999998
last word 0.00115494
same contents 0.001150027
source code 0.001148638
same fashion 0.001142887
same lev 0.001142887
training data 0.00113784
rules 0.00113028
bleu score 0.0011298279999999998
new method 0.001126754
tional cost 0.0011201129999999998
various beam 0.001117419
cost min 0.001109107
source projection 0.001107671
root node 0.001099612
threshold beam 0.001098056
antecedent node 0.001095872
beam settings 0.001094705
cube growing 0.00109417
cube prun 0.001093204
bination cost 0.00108268
antecedent items 0.0010715529999999998
consequent node 0.001057886
other words 0.0010549210000000001
above data 0.001046516
corner item 0.001043037
heuristic function 0.0010421789999999998
costs figure 0.001040184
similar bleu 0.001034047
consequent items 0.001033567
augmented items 0.0010281819999999999
boundary words 0.001018526
bleu scores 0.001017818
consequent item 0.0010162419999999999
tecedent items 0.0010128799999999999
popped items 0.00100911
enumerated items 0.00100911
merate items 0.00100911
first iteration 0.0010063569999999998
sulting item 0.0010032589999999998
