A hotly debated topic in machine translation is human evaluation. On the one hand, it is extremely costly and time consuming; on the other, it is an important and unfortunately inevitable part of any system. This paper describes COSTA MT Evaluation Tool, an open stand-alone tool for human machine translation evaluation. It is a Java program that can be used to manually evaluate the quality of the machine translation output. It is simple in use, designed to allow machine translation potential users and developers to analyze their systems using a friendly environment. It enables the ranking of the quality of machine translation output segment-bysegment for a particular language pair. The benefits of this tool are multiple. Firstly, it is a rich repository of commonly used industry criteria (fluency, adequacy and translation error classification). Secondly, it is freely available to anyone and provides results that can be further analyzed. Thirdly, it estimates the time needed for each evaluated sentence. Finally, it gives suggestions about the fuzzy matching of the candidate translations.
[1]
Joseph Olive,et al.
Handbook of Natural Language Processing and Machine Translation: DARPA Global Autonomous Language Exploitation
,
2011
.
[2]
Christian Federmann.
Appraise: An Open-Source Toolkit for Manual Evaluation of Machine Translation Output
,
2012
.
[3]
Philip Koehn,et al.
Statistical Machine Translation
,
2010,
EAMT.
[4]
Lucia Specia,et al.
PET: a Tool for Post-editing and Assessing Machine Translation
,
2012,
LREC.
[5]
Philipp Koehn,et al.
Moses: Open Source Toolkit for Statistical Machine Translation
,
2007,
ACL.