human evaluation 0.00366037
evaluation metric 0.00303446
evaluation metrics 0.0029717700000000003
reference translation 0.002921872
translation sentence 0.002811172
system translation 0.0027661779999999998
automatic evaluation 0.002720513
translation task 0.002508821
machine translation 0.002440725
translation output 0.002421979
translation tasks 0.002376839
translation texts 0.0023546269999999998
word matches 0.002347391
entire translation 0.002328979
various translation 0.00232203
evaluation datasets 0.0023208309999999998
evaluation criteria 0.002314876
evaluation exercise 0.002270768
ence translation 0.0022703949999999997
erence translation 0.002266097
single word 0.002062773
word alignments 0.002042763
word alignment 0.002038499
human judge 0.00203179
exact word 0.002020402
human judges 0.002001194
human judgement 0.001998567
human judgements 0.001997646
evaluation 0.00197087
translation 0.00195694
semantic role 0.001829263
automatic metric 0.001813233
bleu bleu 0.001795506
bleu score 0.001794956
other metrics 0.0017845740000000001
score scores 0.001776713
automatic metrics 0.0017505430000000002
good correlation 0.001720315
reference translations 0.001706731
rank correlation 0.0016999150000000002
average correlation 0.001659527
recall metric 0.0016447990000000002
other sentence 0.0016379060000000002
similarity metric 0.0016227810000000002
correlation results 0.0016131460000000002
different translations 0.001606195
correlation value 0.0015978540000000001
metric adequacy 0.001588398
precision score 0.001565538
system translations 0.001551037
matching problem 0.001550291
different matches 0.001544747
correlation coefficient 0.0015152610000000002
reference text 0.001495242
reference unigram 0.001494675
unigram matching 0.00149167
source sentence 0.0014906189999999999
recall score 0.001478412
standard metric 0.0014661750000000001
different information 0.001461399
similarity score 0.001456394
semantic roles 0.001455647
corresponding reference 0.001450257
commentary metric 0.0014411740000000001
unigram score 0.0014269460000000001
first reference 0.001426539
matching items 0.001422709
metric design 0.001422447
europarl metric 0.001414356
second reference 0.001411856
news commentary 0.001408334
bipartite matching 0.001407173
role information 0.001399146
reference corpus 0.0013879300000000001
different weights 0.0013867010000000002
metric adq 0.0013855830000000001
matching pairs 0.001384011
weight matching 0.00138243
reference unigrams 0.0013741250000000001
current metrics 0.0013634110000000001
dependency scores 0.001359054
teor metrics 0.001342501
system unigram 0.001338981
reference transla 0.0013374160000000001
news com 0.0013297439999999999
graph matching 0.001307147
matching criteria 0.001305933
reference nodes 0.001305916
reference bigram 0.00128304
matching crosses 0.001280253
reference item 0.001278386
reference rela 0.001276713
sentence pairs 0.001276316
optimal matching 0.001271283
first system 0.001270845
matching conditions 0.00125639
second system 0.001256162
system relations 0.001251862
different alignments 0.001240119
sentence pair 0.001238292
