Skip to Main Content
Table A3 
Spearman correlation with human judgments from WMT16 Metrics Task on the outputs of different MT systems in the WMT16 Direct Assessment data set. This table corresponds to Table 8 in the main body of the article.
PBMTSYNTAXNMT
-TERp-A 0.504 0.509 0.623 
Meteor 0.486 0.500 0.613 
MPEDA 0.478 0.493 0.608 
ROUGE-SU* 0.494 0.458 0.579 
ChrF3 0.478 0.405 0.568 
NIST-4 0.431 0.413 0.548 
BLEU-4 0.388 0.392 0.497 
-WER 0.403 0.379 0.547 
-TER 0.378 0.382 0.549 
-PER 0.382 0.396 0.523 
 
UPF-Cobalt 0.500 0.511 0.551 
CP-Oc(*) 0.407 0.450 0.572 
SP-lNIST 0.453 0.423 0.546 
DP-Oc(*) 0.414 0.363 0.404 
SR-Or(*) 0.316 0.366 0.292 
 
DPMFcomb 0.521 0.555 0.624 
Metrics-F 0.526 0.550 0.620 
Cobalt-F-comp 0.531 0.580 0.592 
UoW-ReVal 0.521 0.535 0.542 
BEER 0.465 0.400 0.579 
PBMTSYNTAXNMT
-TERp-A 0.504 0.509 0.623 
Meteor 0.486 0.500 0.613 
MPEDA 0.478 0.493 0.608 
ROUGE-SU* 0.494 0.458 0.579 
ChrF3 0.478 0.405 0.568 
NIST-4 0.431 0.413 0.548 
BLEU-4 0.388 0.392 0.497 
-WER 0.403 0.379 0.547 
-TER 0.378 0.382 0.549 
-PER 0.382 0.396 0.523 
 
UPF-Cobalt 0.500 0.511 0.551 
CP-Oc(*) 0.407 0.450 0.572 
SP-lNIST 0.453 0.423 0.546 
DP-Oc(*) 0.414 0.363 0.404 
SR-Or(*) 0.316 0.366 0.292 
 
DPMFcomb 0.521 0.555 0.624 
Metrics-F 0.526 0.550 0.620 
Cobalt-F-comp 0.531 0.580 0.592 
UoW-ReVal 0.521 0.535 0.542 
BEER 0.465 0.400 0.579 
Close Modal

or Create an Account

Close Modal
Close Modal