translation system 0.00496569
translation systems 0.00356583
experimental system 0.003407743
metric system 0.003293921
system performance 0.0032841750000000003
baseline system 0.003268869
first system 0.0032165650000000002
system outputs 0.003160183
second system 0.003122245
system pair 0.003119016
wmt system 0.0030485490000000002
system differences 0.003042
hierarchical system 0.003009859
system avg 0.0030029510000000002
system configuration 0.002996532
median system 0.00299374
translation results 0.002938144
machine translation 0.002916706
translation metric 0.002840911
translation quality 0.00274673
translation metrics 0.00274511
system 0.00270935
test data 0.00270128
chine translation 0.00254564
test set 0.002435123
different bleu 0.0023991200000000002
test corpus 0.002288685
translation 0.00225634
english systems 0.002030889
single test 0.001999972
multiple test 0.001973152
training data 0.0019693389999999996
test sets 0.001964057
word segmentation 0.001943614
second test 0.001939225
good test 0.0019376950000000001
crf word 0.001935025
bleu score 0.0019169220000000002
virtual test 0.0018791020000000001
different optimization 0.00186091
permutation test 0.001816768
nonparametric test 0.0018107190000000001
bleu difference 0.001713108
different values 0.0016444720000000001
bleu scores 0.0016388750000000001
model parameters 0.001637609
ample systems 0.001594245
new feature 0.001576129
segmentation model 0.001573004
different degrees 0.001570147
function bleu 0.0015696030000000001
development data 0.001566596
set evaluation 0.001471566
unseen data 0.0014602279999999998
same sentence 0.001457938
feature inde 0.001436602
set scores 0.001433128
metric score 0.001386953
single set 0.001382435
experimental results 0.001380197
set effects 0.001364427
score improvements 0.001361485
tuning set 0.001334898
systems 0.00130949
development set 0.001300439
parallel training 0.001282271
model 0.00127789
statistical machine 0.001277883
large number 0.001243182
other optimization 0.001227479
set spread 0.001193282
many number 0.001188366
output translations 0.001178054
unsegmented words 0.001170767
language pair 0.001170233
many problem 0.001160029
human analysis 0.001156332
feature 0.00114967
metric improvements 0.001143674
rate training 0.001137726
significant results 0.0011277330000000001
metric scores 0.001108906
btec corpus 0.001108417
same units 0.001105651
confidence interval 0.001104234
score stest 0.001103562
ric score 0.001098796
prior work 0.001089634
statistical hypothesis 0.001089595
english ranking 0.001089592
score deltas 0.00108735
confidence estimation 0.001081659
training examples 0.001079464
natural language 0.0010709509999999999
optimizer outputs 0.001061798
experimental result 0.001060831
multiple optimizer 0.001057787
small number 0.00105698
experimental manipulation 0.0010558499999999999
experimental samples 0.0010548559999999998
