LeBLEU: N-gram-based Translation Evaluation Score for Morphologically Complex Languages
Article Status
Published
Authors/contributors
- Virpioja, Sami (Author)
- Grönroos, Stig-Arne (Author)
Title
LeBLEU: N-gram-based Translation Evaluation Score for Morphologically Complex Languages
Date
2015
Proceedings Title
Proceedings of the Tenth Workshop on Statistical Machine Translation
Conference Name
Proceedings of the Tenth Workshop on Statistical Machine Translation
Place
Lisbon, Portugal
Publisher
Association for Computational Linguistics
Pages
411-416
Language
en
Short Title
LeBLEU
Accessed
27/10/2023, 16:48
Library Catalogue
DOI.org (Crossref)
Extra
<AI Smry>: The results on WMT data sets show that fuzzy n-gram matching improves correlations to human evaluation especially for highly compounding languages.
Citation
Virpioja, S., & Grönroos, S.-A. (2015). LeBLEU: N-gram-based Translation Evaluation Score for Morphologically Complex Languages. Proceedings of the Tenth Workshop on Statistical Machine Translation, 411–416. https://doi.org/10.18653/v1/W15-3052
Technical methods
Link to this record