Skip to Main Content
Table 6 
Average overlap between the estimated distributions of the metric scores for each quality level in the WMT17 Quality Estimation data set.
MetricQ1Q2Q3Q4
ROUGE-SU* 0.533 0.685 0.686 0.645 
-TERp-A 0.526 0.661 0.660 0.575 
Meteor 0.561 0.704 0.698 0.605 
ChrF3 0.562 0.705 0.705 0.621 
BLEU-4 0.580 0.728 0.715 0.670 
NIST-4 0.594 0.716 0.696 0.627 
-WER 0.561 0.725 0.735 0.689 
-TER 0.560 0.723 0.731 0.686 
-PER 0.570 0.698 0.713 0.629 
 
UPF-Cobalt 0.533 0.686 0.683 0.595 
DP-Oc(*) 0.546 0.719 0.702 0.689 
CP-Oc(*) 0.557 0.697 0.684 0.610 
SP-lNIST 0.605 0.722 0.702 0.641 
 
BEER 0.582 0.712 0.709 0.636 
POSTECH 0.405 0.586 0.586 0.560 
HTER 0.251 0.503 0.529 0.490 
MetricQ1Q2Q3Q4
ROUGE-SU* 0.533 0.685 0.686 0.645 
-TERp-A 0.526 0.661 0.660 0.575 
Meteor 0.561 0.704 0.698 0.605 
ChrF3 0.562 0.705 0.705 0.621 
BLEU-4 0.580 0.728 0.715 0.670 
NIST-4 0.594 0.716 0.696 0.627 
-WER 0.561 0.725 0.735 0.689 
-TER 0.560 0.723 0.731 0.686 
-PER 0.570 0.698 0.713 0.629 
 
UPF-Cobalt 0.533 0.686 0.683 0.595 
DP-Oc(*) 0.546 0.719 0.702 0.689 
CP-Oc(*) 0.557 0.697 0.684 0.610 
SP-lNIST 0.605 0.722 0.702 0.641 
 
BEER 0.582 0.712 0.709 0.636 
POSTECH 0.405 0.586 0.586 0.560 
HTER 0.251 0.503 0.529 0.490 
Close Modal

or Create an Account

Close Modal
Close Modal