human translation 0.0040148399999999996
translation words 0.0037051899999999997
translation evaluation 0.0035887999999999996
human evaluation 0.00318324
human translations 0.0030406699999999997
reference translation 0.0030122409999999997
candidate translation 0.0029149249999999996
good translation 0.002889306
different word 0.002850569
translation system 0.0028004379999999997
machine translation 0.002788055
measure translation 0.002753582
reference word 0.002745881
translation length 0.0027365099999999997
translation corpus 0.002680218
candidate word 0.0026485650000000003
translation pairs 0.00264591
human reference 0.002606681
translation performance 0.002557247
bad translation 0.0025476699999999997
proper translation 0.0025361629999999997
word error 0.002530182
bleu evaluation 0.0025118099999999997
chine translation 0.002508603
evaluation metric 0.00250844
word order 0.002467055
source word 0.0024366400000000003
human transla 0.002358323
reference words 0.002297031
word choice 0.0022700010000000002
bleu metric 0.00226305
cessful word 0.0022406880000000002
human translators 0.002231888
human labor 0.0022174689999999997
human judgments 0.0022134349999999997
translation 0.0022102
candidate words 0.002199715
human translator 0.0021724179999999997
many words 0.002161208
human evaluations 0.002137629
same words 0.002134061
human judgment 0.002123015
human eval 0.002120097
human judg 0.002115096
human judges 0.002109424
other words 0.0021088360000000002
distinguishing human 0.002104537
human evalu 0.002100636
bleu score 0.00209023
reference translations 0.002038071
bleu scores 0.001991439
automatic evaluation 0.001975334
evaluation system 0.001968838
candidate translations 0.001940755
good translations 0.001915136
important words 0.0018929090000000001
english translations 0.0018634139999999999
machine translations 0.0018138849999999999
evaluation figure 0.001804873
human 0.00180464
synonymous words 0.001794484
spurious words 0.0017927630000000002
different language 0.001792072
evaluation experiment 0.001724857
translations ranking 0.0016877419999999999
evaluation techniques 0.0016801960000000001
precision score 0.0016783140000000002
evaluation bottleneck 0.001674531
average score 0.001626721
sentence test 0.001598915
precision scores 0.001579523
bad translations 0.0015734999999999998
rate metric 0.001562539
log bleu 0.001555524
ence translations 0.001548293
translations accord 0.001535827
erence translations 0.001534779
number metric 0.001527368
baseline bleu 0.001520806
baseline metric 0.001517436
words 0.00149499
metric performance 0.001476887
many reference 0.001468259
score difference 0.001431378
bleu imple 0.001429652
bleu details 0.001429652
numerical metric 0.001428477
same sentence 0.001427673
closeness metric 0.001426955
metric ranges 0.001426955
correlation coefficient 0.001394586
average precision 0.001390995
good candidate 0.0013838309999999999
evaluation 0.0013786
sentence level 0.001363242
reference transla 0.001355724
high correlation 0.001353555
candidate sentences 0.001345457
target language 0.0013359280000000001
reference length 0.001328351
