language model 0.00392692
source word 0.00377875
model score 0.003310944
source phrase 0.0032553
target phrase 0.00318156
model state 0.003071789
distortion model 0.003062949
model features 0.002928688
translation table 0.002884791
translation hypotheses 0.0028834069999999997
machine translation 0.002849739
model cpu 0.002821328
model scores 0.002762019
guage model 0.002744621
translation systems 0.002723017
model likes 0.002716795
tween model 0.002716211
translation speed 0.0026810619999999997
translation rules 0.002677013
translation hypothe 0.002663164
modern translation 0.002653441
chine translation 0.002650439
phrase table 0.002648791
idiom translation 0.002647843
phrase pair 0.0026471940000000003
standard phrase 0.002608169
source words 0.00256621
memory phrase 0.0025632460000000004
text phrase 0.002539561
phrase trie 0.0025153500000000004
target words 0.00249247
present phrase 0.0024634830000000003
model 0.00243506
phrase tables 0.002415107
phrase tries 0.002411865
translation 0.002368
language models 0.002310261
phrase 0.002132
source sentence 0.0021090550000000003
specific words 0.001964011
multiple words 0.00186758
few words 0.001801086
boundary words 0.0017958849999999999
source phrases 0.0017760710000000002
oped language 0.001771412
common words 0.0017581909999999999
words cat 0.001744336
suffix words 0.0017266
discontiguous words 0.001723591
target phrases 0.001702331
source length 0.0015313380000000001
total source 0.0015216700000000001
language 0.00149186
multiple target 0.00147423
source sen 0.001466373
words 0.00144291
compatible source 0.001426082
source coverage 0.001423092
source cov 0.001404497
source cover 0.0014029490000000001
same search 0.001401944
target lan 0.00139242
sentence pairs 0.001388349
new decoder 0.0013775760000000001
group target 0.0013405959999999999
new search 0.001327889
parallel data 0.001312484
trie data 0.001306738
same state 0.001296729
maximum score 0.0012843059999999998
hypothesis state 0.001255361
search space 0.00125328
cpu bleu 0.001251577
detailed models 0.001245422
forum data 0.001219001
additional state 0.001215454
search algorithm 0.001213958
search accuracy 0.001208584
data structure 0.001207513
tains data 0.001205969
new hypothesis 0.001204577
test set 0.001182002
decoding algorithm 0.001158732
score adjustments 0.001157565
score adjustment 0.001157565
search process 0.001153193
bleu stack 0.001151809
peak bleu 0.0011470859999999999
uncased bleu 0.0011470859999999999
moses decoder 0.001138197
same accuracy 0.00112664
distortion feature 0.001126254
source 0.0011233
pruning pass 0.001120594
decoder reasons 0.001116649
other hypothesis 0.001114521
beam search 0.001103403
same number 0.001102656
new hypotheses 0.0011013519999999999
refining models 0.001098414
