Elamur Research⌕
☰
home / papers / 2026 / The MaTOS Pipeline for the Translation of Scientific Abstracts on the HAL Platform

The MaTOS Pipeline for the Translation of Scientific Abstracts on the HAL Platform

Summary

English-language dominance in academic publishing disadvantages researchers whose first language is not English. Within the ANR-funded MaTOS project, the authors built an automatic pipeline that increases the number of bilingual English/French abstracts on HAL, the French national open archive. It runs in five stages:

  • daily metadata harvesting through the HAL API;
  • language checking and cleaning of source abstracts;
  • translation with EuroLLM-9B-Instruct;
  • filtering of weak outputs;
  • sending drafts to authors, who approve or post-edit them.

Abstracts count as metadata, so they can be translated even when an article lacks an open licence. The authors compared three input granularities: single sentences, three-sentence chunks, and full abstracts. They also tested zero-, one- and two-shot prompting, scoring the outputs with the reference-free metrics CometKiwi and GEMBA-MQM. CometKiwi gave a slight edge to shorter segments, but GEMBA-MQM, applied to whole abstracts, ranked sentence-by-sentence output lowest. On that basis the team adopted document-level translation. The released code and the author-validated translations should support multilingual open science and provide parallel data for MT research.

Key points

  • Scale of the gap (February 2026): HAL holds 4.5M references. About 1.7M have only an English abstract, and roughly 582k have only a French one.
  • Source-side filtering:
    • fastText checks that the declared language of each abstract matches its actual language.
    • Abstracts under 15 words, or ending in an ellipsis, are discarded.
    • HTML tags are stripped, and sub- and superscripts are normalised.
  • Translation setup:
    • EuroLLM-9B-Instruct runs through vLLM within a 4096-token context limit.
    • Each request takes about 5.95 s on an A100 GPU.
    • Few-shot examples come from the ACAD-TRAIN and ACAD-BENCH subsets of ACADATA. They are chosen by BM25 over text lemmatised with SpaCy.
  • Effect of few-shot prompting: adding examples generally lowered CometKiwi scores. Two-shot prompting also worsened GEMBA-MQM scores, except for fr–en on the HAL-Inria data.
  • Error analysis (10 en–fr abstracts): mistranslation of scientific terms was the most frequent error. Inconsistent terminology increased with shorter segments. Gender and subject-verb agreement errors also appeared. Code is available at github.com/ANR-MaTOS/HAL_pipeline.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages · Machine translation

Cite

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{tsolakis2026matos,
  title = {{The MaTOS Pipeline for the Translation of Scientific Abstracts on the HAL Platform}},
  author = {Tsolakis, Panagiotis and Peng, Ziqian and Romary, Laurent and Yvon, François and Bawden, Rachel},
  booktitle = {Proceedings of the 26th Annual Conference of the European Association for Machine Translation (Volume 2)},
  year = {2026},
  url = {https://aclanthology.org/2026.eamt-2.30/}
}

References

50 references, 20 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Bahdanau, Dzmitry, Kyunghyun Cho, and Yoshua Bengio. 2015. Neural machine translation by jointly learning to align and translate. In Proceedings of the first International Conference on Learning Representations, San Diego, CA. find our card
  2. Bawden, Rachel and François Yvon. 2023. Investigating the translation performance of a large multilingual language model: the case of BLOOM. In Nurminen, Mary, Judith Brenner, Maarit Koponen, Sirkku Latomaa, Mikhail Mikhailov, Frederike Schierl, Tharindu Ranasinghe, Eva Vanmassenhove, Sergi Alvarez… find
  3. Bawden, Rachel, Maud Bénard, Éric de la Clergerie, José Cornejo Cárcamo, Nicolas Dahan, Manon Delorme, Mathilde Huguin, Natalie Kübler, Paul Lerner, Alexandra Mestivier, Joachim Minder, Jean-François Nominé, Ziqian Peng, Laurent Romary, Panagiotis Tsolakis, Lichao Zhu, and François Yvon. 202… find
  4. Bénard, Maud, Alexandra Mestivier, Natalie Kubler, Lichao Zhu, Rachel Bawden, Eric De La Clergerie, Laurent Romary, Mathilde Huguin, Jean-François Nominé, Ziqian Peng, and François Yvon. 2023. MaTOS: Traduction automatique pour la science ouverte. In Boudin, Florian, Béatrice Daille, Richard Du… find
  5. Textes Scientifiques” (ARTS)@TALN 2023, pages 8– 15, Paris, France, 6. ATALA. find
  6. Dahan, Nicolas, Rachel Bawden, and François Yvon. 2026. MetaDocEval: A Contrastive Framework for Evaluating Machine Translation Metrics at the Document-Level. In Proceedings of the 26th Annual Conference of the European Association for Machine Translation, Tilburg, the Netherlands, June. European A… find
  7. Fernandes, Patrick, Kayo Yin, Emmy Liu, André Martins, and Graham Neubig. 2023. When Does Translation Require Context? A Data-driven, Multilingual Exploration. In Rogers, Anna, Jordan Boyd-Graber, and Naoaki Okazaki, editors, Proceedings of the 61st Annual Meeting of the Association for Computation… find
  8. Freitag, Markus, George Foster, David Grangier, Viresh Ratnakar, Qijun Tan, and Wolfgang Macherey. 2021. Experts, errors, and context: A large-scale study of human evaluation for machine translation. Transactions of the Association for Computational Linguistics, 9:1460–1474. find our card
  9. Gordin, Michael D. 2015. Scientific Babel. University of Chicago Press. find
  10. Guo, Jiaxin, Yuanchang Luo, Daimeng Wei, Ling Zhang, Zongyao Li, Hengchao Shang, Zhiqiang Rao, Shaojun Li, Jinlong Yang, Zhanglin Wu, and Hao Yang. 2025. Doc-Guided Sent2Sent++: A Sent2Sent++ Agent with Doc-Guided memory for document-level Machine Translation. arXiv Preprint:2501.08523 [cs], January… — [arXiv:2501.08523](https://arxiv.org/abs/2501.08523) find
  11. Hardmeier, Christian. 2012. Discourse in Statistical Machine Translation: A Survey and a Case Study. Discours - Revue de linguistique, psycholinguistique et informatique, (11). find
  12. Hendy, Amr, Mohamed Abdelrehim, Amr Sharaf, Vikas Raunak, Mohamed Gabr, Hitokazu Matsushita, Young Jin Kim, Mohamed Afify, and Hany Hassan Awadalla. 2023. How good are GPT models at machine translation? A comprehensive evaluation. CoRR, abs/2302.09210. find
  13. Honnibal, Matthew, Ines Montani, Sofie Van Lan-deghem, and Adriane Boyd. 2020. spaCy: Industrial-strength Natural Language Processing in Python. find
  14. Hu, Tianxiang, Pei Zhang, Baosong Yang, Jun Xie, Derek F. Wong, and Rui Wang. 2024. Large language model for multi-domain translation: Benchmarking and domain CoT fine-tuning. In Al-Onaizan, Yaser, Mohit Bansal, and Yun-Nung Chen, editors, Findings of the Association for Computational Linguistics: E… find
  15. Joulin, Armand, Edouard Grave, Piotr Bojanowski, Matthijs Douze, Hervé Jégou, and Tomás Mikolov. 2016. Fasttext.zip: Compressing text classification models. CoRR, abs/1612.03651. find our card
  16. Joulin, Armand, Edouard Grave, Piotr Bojanowski, and Tomas Mikolov. 2017. Bag of tricks for efficient text classification. In Lapata, Mirella, Phil Blunsom, and Alexander Koller, editors, Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Vol… find
  17. Karpinska, Marzena and Mohit Iyyer. 2023. Large language models effectively leverage document-level context for literary translation, but critical errors persist. In Koehn, Philipp, Barry Haddow, Tom Kocmi, and Christof Monz, editors, Proceedings of the Eighth Conference on Machine Translation, page… find our card
  18. Kocmi, Tom and Christian Federmann. 2023a. GEMBA-MQM: Detecting translation quality error spans with GPT-4. In Koehn, Philipp, Barry Haddow, Tom Kocmi, and Christof Monz, editors, Proceedings of the Eighth Conference on Machine Translation, pages 768–775, Singapore, December. Association for Computa… find our card
  19. Kocmi, Tom and Christian Federmann. 2023b. Large language models are state-of-the-art evaluators of translation quality. In Nurminen, Mary, Judith Brenner, Maarit Koponen, Sirkku Latomaa, Mikhail Mikhailov, Frederike Schierl, Tharindu Ranasinghe, Eva Vanmassenhove, Sergi Alvarez Vidal, Nora Aranberr… find
  20. Kwon, Woosuk, Zhuohan Li, Siyuan Zhuang, Ying Sheng, Lianmin Zheng, Cody Hao Yu, Joseph Gonzalez, Hao Zhang, and Ion Stoica. 2023. Efficient memory management for large language model serving with pagedattention. In Proceedings of the 29th Symposium on Operating Systems Principles, SOSP ’23, page 61… find
  21. Lacunza, Iñaki, Javier Garcia Gilabert, Francesca De Luca Fornaciari, Javier Aula-Blasco, Aitor Gonzalez-Agirre, Maite Melero, and Marta Villegas. 2026. ACAData: Parallel dataset of academic data for machine translation. In Piperidis, Stelios, Núria Bel, Henk van den Heuvel, Nancy Ide, Simon Krek,… find
  22. Lambert, Patrik, Jean Senellart, Laurent Romary, Holger Schwenk, Florian Zipser, Patrice Lopez, and Frédéric Blain. 2012. Collaborative machine translation service for scientific texts. In Segond, Frédérique, editor, Proceedings of the Demonstrations at the 13th Conference of the European Chapte… find
  23. Lewis, Mike, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Veselin Stoyanov, and Luke Zettlemoyer. 2020. BART: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension. In Jurafsky, Dan, Joyce Chai, Natalie Schluter,… find our card
  24. Liu, Lei and Min Zhu. 2022. Bertalign: Improved word embedding-based sentence alignment for Chinese–English parallel corpora of literary texts. Digital Scholarship in the Humanities, 38(2):621–634, 12. find
  25. Lommel, Arle Richard, Aljoscha Burchardt, and Hans Uszkoreit. 2013. Multidimensional quality metrics: a flexible system for assessing translation quality. In Proceedings of Translating and the Computer 35, London, UK, November 28-29. Aslib. find our card
  26. Lopes, António, M. Amin Farajian, Rachel Bawden, Michael Zhang, and André F. T. Martins. 2020. Document-level neural MT: A systematic comparison. In Martins, André, Helena Moniz, Sara Fumega, Bruno Martins, Fernando Batista, Luisa Coheur, Carla Parra, Isabel Trancoso, Marco Turchi, Arianna Bisazz… find our card
  27. Lu, Hongyuan, Haoran Yang, Haoyang Huang, Dongdong Zhang, Wai Lam, and Furu Wei. 2024a. Chain-of-dictionary prompting elicits translation in large language models. In Al-Onaizan, Yaser, Mohit Bansal, and Yun-Nung Chen, editors, Proceedings of the 2024 Conference on Empirical Methods in Natural Langu… find
  28. Lu, Qingyu, Baopu Qiu, Liang Ding, Kanjian Zhang, Tom Kocmi, and Dacheng Tao. 2024b. Error analysis prompting enables human-like translation evaluation in large language models. In Ku, Lun-Wei, Andre Martins, and Vivek Srikumar, editors, Findings of the Association for Computational Linguistics: ACL… find our card
  29. Lù, Xing Han. 2024. BM25S: orders of magnitude faster lexical search via eager sparse scoring. find
  30. Martins, Pedro Henrique, João Alves, Patrick Fernandes, Nuno Miguel Guerreiro, Ricardo Rei, M. Amin Farajian, Mateusz Klimaszewski, Duarte M. Alves, José Pombal, Nicolas Boizard, Manuel Faysse, Pierre Colombo, François Yvon, Barry Haddow, José G. C. de Souza, Alexandra Birch, and André F. T. Ma… find
  31. Miculicich, Lesly, Dhananjay Ram, Nikolaos Pappas, and James Henderson. 2018. Document-level neural machine translation with hierarchical attention networks. In Riloff, Ellen, David Chiang, Julia Hockenmaier, and Jun’ichi Tsujii, editors, Proceedings of the 2018 Conference on Empirical Methods in Na… find our card
  32. Moslem, Yasmin, Rejwanul Haque, John D. Kelleher, and Andy Way. 2023. Adaptive machine translation with large language models. In Nurminen, Mary, Judith Brenner, Maarit Koponen, Sirkku Latomaa, Mikhail Mikhailov, Frederike Schierl, Tharindu Ranasinghe, Eva Vanmassenhove, Sergi Alvarez Vidal, Nora Ar… find our card
  33. Nguyen, Minh Van, Viet Dac Lai, Amir Pouran Ben Veyseh, and Thien Huu Nguyen. 2021. Trankit: A light-weight transformer-based toolkit for multilingual natural language processing. In Gkatzia, Dimitra and Djamé Seddah, editors, Proceedings of the 16th Conference of the European Chapter of the Associ… find our card
  34. O’Brien, Dayyán, Bhavitvya Malik, Ona de Gibert, Pinzhen Chen, Barry Haddow, and Jörg Tiedemann. 2025. DocHPLT: A massively multilingual document-level translation dataset. In Proceedings of the Tenth Conference on Machine Translation, pages 286–300, Stroudsburg, PA, USA, September. Association fo… find
  35. Oncevay, Arturo, Charese Smiley, and Xiaomo Liu. 2025. The impact of domain-specific terminology on machine translation for finance in European languages. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Tec… find
  36. Papineni, Kishore, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting on association for computational linguistics, pages 311–318. Association for Computational Linguistics. find our card
  37. Peng, Ziqian, Rachel Bawden, and François Yvon. 2025a. Investigating length issues in document-level machine translation. In Bouillon, Pierrette, Johanna Gerlach, Sabrina Girletti, Lise Volkart, Raphael Rubino, Rico Sennrich, Ana C. Farinha, Marco Gaido, Joke Daems, Dorothy Kenny, Helena Moniz, and… find
  38. Peng, Ziqian, Rachel Bawden, and François Yvon. 2025b. Self-retrieval from distant contexts for document-level machine translation. In Haddow, Barry, Tom Kocmi, Philipp Koehn, and Christof Monz, editors, Proceedings of the Tenth Conference on Machine Translation, pages 220–240, Suzhou, China, Novem… find
  39. Popović, Maja. 2015. chrF: character n-gram F-score for automatic MT evaluation. In Proceedings of the tenth workshop on statistical machine translation, pages 392–395. find our card
  40. Raffel, Colin, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J. Liu. 2020. Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of Machine Learning Research, 21(1). find
  41. Rei, Ricardo, Craig Stewart, Ana C Farinha, and Alon Lavie. 2020. COMET: A neural framework for MT evaluation. In Webber, Bonnie, Trevor Cohn, Yulan He, and Yang Liu, editors, Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 2685–2702, Online, Nov… find our card
  42. Rei, Ricardo, Ana C Farinha, Chrysoula Zerva, Daan van Stigt, Craig Stewart, Pedro Ramos, Taisiya Glushkova, André F. T. Martins, and Alon Lavie. 2021. Are References Really Needed? Unbabel-IST 2021 Submission for the Metrics Shared Task. In Barrault, Loic, Ondrej Bojar, Fethi Bougares, Rajen Chatt… find our card
  43. Rei, Ricardo, Marcos Treviso, Nuno M. Guerreiro, Chrysoula Zerva, Ana C Farinha, Christine Maroti, José G. C. de Souza, Taisiya Glushkova, Duarte Alves, Luisa Coheur, Alon Lavie, and André F. T. Martins. 2022a. find
  44. CometKiwi: IST-unbabel 2022 submission for the quality estimation shared task. In Koehn, Philipp, Loı̈c Barrault, Ondřej Bojar, Fethi Bougares, Rajen Chatterjee, Marta R. Costa-jussà, Christian Federmann, Mark Fishel, Alexander Fraser, Markus Freitag, Yvette Graham, Roman Grundkiewicz, Paco Guzman… find our card
  45. Rei, Ricardo, Marcos Treviso, Nuno M. Guerreiro, Chrysoula Zerva, Ana C Farinha, Christine Maroti, José G. C. de Souza, Taisiya Glushkova, Duarte Alves, Luisa Coheur, Alon Lavie, and André F. T. Martins. 2022b. find
  46. CometKiwi: IST-unbabel 2022 submission for the quality estimation shared task. In Koehn, Philipp, Loı̈c Barrault, Ondřej Bojar, Fethi Bougares, Rajen Chatterjee, Marta R. Costa-jussà, Christian Federmann, Mark Fishel, Alexander Fraser, Markus Freitag, Yvette Graham, Roman Grundkiewicz, Paco Guzman… find our card
  47. Tan, Weiting, Haoran Xu, Lingfeng Shen, Shuyue Stella Li, Kenton Murray, Philipp Koehn, Benjamin Van Durme, and Yunmo Chen. 2024. Narrowing the gap between zero- and few-shot machine translation by matching styles. In Duh, Kevin, Helena Gomez, and Steven Bethard, editors, Findings of the Association… find
  48. Wang, Longyue, Chenyang Lyu, Tianbo Ji, Zhirui Zhang, Dian Yu, Shuming Shi, and Zhaopeng Tu. 2023. Document-level machine translation with large language models. In Bouamor, Houda, Juan Pino, and Kalika Bali, editors, Proceedings of the 2023 Conference on Empirical Methods in Natural Language Proces… find our card
  49. Xiao, Tong, Jingbo Zhu, Shujie Yao, and Hao Zhang. 2011. Document-level consistency verification in machine translation. In Proceedings of Machine Translation Summit XIII: Papers, Xiamen, China, September 19-23. find
  50. Zhu, Wenhao, Hongyi Liu, Qingxiu Dong, Jingjing Xu, Shujian Huang, Lingpeng Kong, Jiajun Chen, and Lei Li. 2024. Multilingual machine translation with large language models: Empirical results and analysis. In Duh, Kevin, Helena Gomez, and Steven Bethard, editors, Findings of the Association for Comp… find our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/tsolakis-2026-matos-pipeline-translation-scientific-abstracts/