human translation 0.0035757
evaluation metric 0.0034369099999999996
human translations 0.0032842799999999997
evaluation metrics 0.00326697
human reference 0.00307259
automatic evaluation 0.0027525099999999997
human references 0.002718566
human judgments 0.002674951
automatic metrics 0.00255742
human effort 0.00252931
human judges 0.0024953149999999997
human judg 0.002488289
human translators 0.0024750939999999997
human involvement 0.002471964
evaluation task 0.002356535
evaluation references 0.002341496
translation quality 0.002200984
reference translations 0.00214067
standard metrics 0.002106228
trained metric 0.002106126
composite evaluation 0.002094733
evaluation loop 0.002094733
translation fluency 0.00209227
referenceless metric 0.002072129
trained metrics 0.001936186
individual metrics 0.00192921
translation sen 0.0018490999999999998
chine translation 0.0018377509999999999
good translations 0.001835947
language model 0.001781931
evaluation 0.00173103
metric 0.00170588
different features 0.00168926
same source 0.001669842
different learning 0.001574868
model features 0.0015632810000000001
adequate translations 0.001553573
imperfect translations 0.001547568
metrics 0.00153594
standard reference 0.001534778
training data 0.001531939
automatic assessment 0.0015169860000000001
research systems 0.001493373
translation 0.0014676
single reference 0.001447286
input sentence 0.0014242360000000002
new test 0.001361656
ing model 0.001326596
large text 0.00131694
test instances 0.001309427
feature vector 0.001306261
such instance 0.00125062
sourced research 0.001242511
put sentence 0.0012326680000000001
linguistic indicators 0.001231842
final score 0.001224077
translations 0.00117618
ing framework 0.001170457
main findings 0.001143759
main insight 0.001143759
large body 0.0011227
large quantity 0.0011227
quality assurance 0.0011004650000000001
useful information 0.001098469
multiple ref 0.001080161
ence sentences 0.001043078
training examples 0.001033228
automatic 0.00102148
fluency fea 0.0010194190000000001
regression objective 0.001019257
learning algorithm 0.0010177670000000002
man references 0.001017254
gold standard 0.00101725
text corpora 0.001012823
pseudo references 0.0010116320000000002
alternative approach 0.0010108650000000001
subjective task 9.98881E-4
learning regime 9.91059E-4
parallel corpora 9.69307E-4
objective question 9.68366E-4
experimental results 9.64826E-4
reference 9.6449E-4
important factor 9.59423E-4
language 9.5624E-4
ency judgments 9.32572E-4
challenging problem 9.317489999999999E-4
related work 9.300280000000001E-4
initial results 9.18722E-4
feature 9.09624E-4
pseudo ref 9.00967E-4
nuanced comparisons 8.97219E-4
empirical studies 8.873290000000001E-4
source 8.80888E-4
ison points 8.76109E-4
previous datasets 8.7584E-4
poor outputs 8.728270000000001E-4
significant expenditure 8.64235E-4
sentence 8.62743E-4
fluent outputs 8.61792E-4
long period 8.58003E-4
