Skip to main navigation Skip to search Skip to main content

Extending BLEU evaluation method with linguistic weight

  • Muyun Yang*
  • , Junguo Zhu
  • , Jufeng Li
  • , Lixin Wang
  • , Haoliang Qi
  • , Sheng Li
  • , Liu Daxin
  • *Corresponding author for this work
  • Harbin Engineering University
  • School of Computer Science and Technology, Harbin Institute of Technology
  • Harbin Institute of Technology
  • Heilongjiang Institute of Technology

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

BLEU is one of the most popular metrics for automatic evaluation of machine translation quality. Focusing on its ignorance of different effects of various translation units upon translation quality, this paper extends proper weights to different words and n-grams in the framework of BLEU. The linear regression method is adopted to capture the human perception on translation quality via word types and n-gram length. Compared with other linguistic-rich metrics based on machine learning, the proposed approach is simple and largely preserves BLEU's advantage of language independence. Experimental results indicate that this method brings a much better evaluation performance for both human translation and machine translation than original BLEU.

Original languageEnglish
Title of host publicationProceedings of the 9th International Conference for Young Computer Scientists, ICYCS 2008
Pages1683-1688
Number of pages6
DOIs
StatePublished - 2008
Externally publishedYes
Event9th International Conference for Young Computer Scientists, ICYCS 2008 - Zhang Jia Jie, Hunan, China
Duration: 18 Nov 200821 Nov 2008

Publication series

NameProceedings of the 9th International Conference for Young Computer Scientists, ICYCS 2008

Conference

Conference9th International Conference for Young Computer Scientists, ICYCS 2008
Country/TerritoryChina
CityZhang Jia Jie, Hunan
Period18/11/0821/11/08

Keywords

  • BLEU
  • Linear regression
  • Linguistic weight
  • Machine translation evaluation

Fingerprint

Dive into the research topics of 'Extending BLEU evaluation method with linguistic weight'. Together they form a unique fingerprint.

Cite this