ŠTEFÁNIK, Michal, Vít NOVOTNÝ and Petr SOJKA. Regressive Ensemble for Machine Translation Quality Evaluation. Online. In Loïc Barrault et al. Proceedings of EMNLP 2021 Sixth Conference on Machine Translation (WMT 21). Online and Punta Cana, Dominican Republi: ACL, 2021. p. 1041-1048. ISBN 978-1-954085-94-7. [citováno 2024-04-23]
Other formats:   BibTeX LaTeX RIS
Basic information
Original name Regressive Ensemble for Machine Translation Quality Evaluation
Authors ŠTEFÁNIK, Michal (703 Slovakia, belonging to the institution), Vít NOVOTNÝ (203 Czech Republic, belonging to the institution) and Petr SOJKA (203 Czech Republic, guarantor, belonging to the institution)
Edition Online and Punta Cana, Dominican Republi, Proceedings of EMNLP 2021 Sixth Conference on Machine Translation (WMT 21), p. 1041-1048, 8 pp. 2021.
Publisher ACL
Other information
Original language English
Type of outcome Proceedings paper
Field of Study 60203 Linguistics
Country of publisher United States of America
Confidentiality degree is not subject to a state or trade secret
Publication form electronic version available online
WWW preprint paper
RIV identification code RIV/00216224:14330/21:00122292
Organization unit Faculty of Informatics
ISBN 978-1-954085-94-7
Keywords (in Czech) strojový překlad; automatické vyhodnocení kvality překladu;
Keywords in English machine translation; translation quality metrics; regressive ensemble for machine translation quality evaluation
Tags firank_B
Tags International impact, Reviewed
Changed by Changed by: RNDr. Pavel Šmerk, Ph.D., učo 3880. Changed: 23/5/2022 14:57.
Abstract
This work introduces a simple regressive ensemble for evaluating machine translation quality based on a set of novel and established metrics. We evaluate the ensemble using a correlation to expert-based MQM scores of the WMT 2021 Metrics workshop. In both monolingual and zero-shot cross-lingual settings, we show a significant performance improvements over single systems. In the cross-lingual settings, we also demonstrate that an ensemble approach is well-applicable to unseen languages. Furthermore, we identify a strong reference-free baseline that consistently outperforms the commonly-used BLEU and METEOR measures and significantly improves our ensemble's performance.
Links
MUNI/A/1549/2020, interní kód MUName: Zapojení studentů Fakulty informatiky do mezinárodní vědecké komunity 21 (Acronym: SKOMU)
Investor: Masaryk University
PrintDisplayed: 23/4/2024 21:29