Elamur Research⌕
☰
home / papers / 2026 / Bi-Text Mining across German Dialects: On the Role of Synthetic Training Data for Dialect Adaptation

Bi-Text Mining across German Dialects: On the Role of Synthetic Training Data for Dialect Adaptation

Paper · 2026 · Jing Wang, Barbara Plank and Robert Litschko

Summary

Multilingual sentence encoders perform well on standardized languages, but their behavior on dialects has received little attention. Dialects sit on a continuum with the standard language. They share much of its vocabulary but differ in spelling, word choice and idioms. The authors test how well four widely used bi-encoders can retrieve Standard German translations for queries in Low German, Bavarian and Alemannic, using English–German as a high-resource reference. They checked the available parallel resources and found the web-derived corpora unreliable, so they built the benchmark from Tatoeba.

Parallel dialect data is scarce, so the study also tests synthetic data as weak supervision for adapting the models. Two sources are used: word-level dictionary translations and GPT-4o translations. Without adaptation, the encoders struggle to match dialect sentences with their German equivalents. Fine-tuning on synthetic pairs improves retrieval considerably, and LLM-generated data gives the larger gains. A controlled analysis on Bavarian shows that performance falls once dialect words exceed about 60% of a text.

The work offers a practical recipe for dialect–standard bi-text mining when real data is scarce. The released code and data are useful for low-resource variety work.

Key points

  • Setup: Translation retrieval on Tatoeba with 1K dialect queries per language pair (nds, bar, als, plus en as reference) against a shared pool of 100K German documents. The models include LaBSE, GTE, BGE-M3 and Qwen3.

  • Data quality: The authors manually checked 1,000 sampled pairs per language from the web-derived corpora. Misalignment reached 33.2% for nds-de and 47.7% for bar-de, and 28.9% of Bavarian-side entries were actually Standard German. BM25 also showed strong lexical shortcuts, so WikiMatrix and Wikimedia were excluded from evaluation.

  • Synthetic data bias: Only 90 authentic Bavarian pairs exist in Tatoeba, so 910 of the 1,000 Bavarian queries were produced by GPT-4o. On the 90 human-translated pairs, BGE-M3 reaches the following P@1 scores:

    • 0.219 zero-shot
    • 0.329 after dictionary fine-tuning
    • 0.685 after LLM fine-tuning

    The same LLM-fine-tuned model scores 0.781 on GPT-4o versions of those pairs. This suggests that evaluating on LLM-translated text somewhat inflates the scores.

  • Code-mixing robustness: All models retrieve reliably while dialect words make up 60% or less of the text, and effectiveness drops beyond that.

  • Resources: Code and data are available at github.com/mainlp/dialect-bitext-mining.

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

 Institutions

🎓Ludwig-Maximilians-Universität MünchenMunich · Germany107 works · 56,478 citations🎓Munich Center for Machine LearningMunich · Germany57 works · 1,094 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksInformation Systems · 1,437 works

Cite

Cite as Jing Wang, Barbara Plank and Robert Litschko. 2026. Bi-Text Mining across German Dialects: On the Role of Synthetic Training Data for Dialect Adaptation. Proceedings of the 19th Workshop on Building and Using Comparable Corpora (BUCC).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/3gmqhegz45cn

@inproceedings{wang2026bi,
  title = {{Bi-Text Mining across German Dialects: On the Role of Synthetic Training Data for Dialect Adaptation}},
  author = {Jing Wang, Barbara Plank and Robert Litschko},
  booktitle = {Proceedings of the 19th Workshop on Building and Using Comparable Corpora (BUCC)},
  year = {2026},
  doi = {10.63317/3gmqhegz45cn},
  url = {https://doi.org/10.63317/3gmqhegz45cn}
}

References

53 references, 19 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Noëmi Aepli, Chantal Amrhein, Florian Schottmann, and Rico Sennrich. 2023. A benchmark for evaluating machine translation metrics on dialects without standard orthography. In Proceedings of the Eighth Conference on Machine Translation, pages 1045–1065, Singapore. Association for Computational Linguistics. find
  2. Md Mahfuz Ibn Alam, Sina Ahmadi, and Antonios Anastasopoulos. 2024. A morphologically-aware dictionary-based data augmentation technique for machine translation of under-represented languages. ↗
  3. Ekaterina Artemova and Barbara Plank. 2023. Low-resource bilingual dialect lexicon induction with large language models. In Proceedings of the 24th Nordic Conference on Computational Linguistics (NoDaLiDa), pages 371–385, Tórshavn, Faroe Islands. University of Tartu Library. ↗
  4. Mikel Artetxe and Holger Schwenk. 2019. Margin-based parallel corpus mining with multilingual sentence embeddings. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3197–3203, Florence, Italy. Association for Computational Linguistics. ↗
  5. Adam L. Berger, Rich Caruana, David A. Cohn, Dayne Freitag, and Vibhu Mittal. 2000. Bridging the lexical chasm: statistical approaches to answer-finding. In Annual International ACM SI-GIR Conference on Research and Development in Information Retrieval. find
  6. Verena Blaschke, Barbara Kovačić, Siyao Peng, Hinrich Schütze, and Barbara Plank. 2024. MaiBaam: A multi-dialectal Bavarian Universal Dependency treebank. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pages 10921– 10938, Torino, Italia. ELRA and ICCL. find
  7. Verena Blaschke, Hinrich Schuetze, and Barbara Plank. 2023. A survey of corpora for Germanic low-resource languages and dialects. In Proceedings of the 24th Nordic Conference on Computational Linguistics (NoDaLiDa), pages 392– 414, Tórshavn, Faroe Islands. University of Tartu Library. ↗
  8. Minh Duc Bui, Manuel Mager, Peter Herbert Kann, and Katharina von der Wense. 2026. Meenz bleibt meenz, but large language models do not speak its dialect. ↗
  9. Manuel Burghardt, Daniel Granvogl, and Christian Wolff. 2016. Creating a lexicon of Bavarian dialect by means of Facebook language data and crowdsourcing. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC’16), pages 2029– 2033, Portorož, Slovenia. European Language Resources Association (ELRA). find
  10. Jianlyu Chen, Shitao Xiao, Peitian Zhang, Kun Luo, Defu Lian, and Zheng Liu. 2024. M3-embedding: Multi-linguality, multi-functionality, multi-granularity text embeddings through self-knowledge distillation. In Findings of the Association for Computational Linguistics: ACL 2024, pages 2318–2335, Bangkok, Thailand. Association for Computational Linguistics.
  11. Christian Chiarcos, Janine Siewert, Tabea Gröger, and Christian Fäth. 2025. Towards a cross-dialectal dictionary for Low German (Low Saxon). In Proceedings of the 21st Conference on Natural Language Processing (KONVENS 2025): Long and Short Papers, pages 282–294, Hannover, Germany. HsH Applied Academics. find
  12. Ona de Gibert, Joseph Attieh, Teemu Vahtola, Mikko Aulamo, Zihao Li, Raúl Vázquez, Tiancheng Hu, and Jörg Tiedemann. 2025. Scaling low-resource MT via synthetic data generation with LLMs. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 27674–27692, Suzhou, China. Association for Computational Linguistics. ↗
  13. Daniel Deutsch, Eleftheria Briakou, Isaac Caswell, Mara Finkelstein, Rebecca Galor, Juraj Juraska, Geza Kovacs, Alison Lui, Ricardo Rei, Jason Riesa, Shruti Rijhwani, Parker Riley, Elizabeth Salesky, Firas Trabelsi, Stephanie Winkler, Biao Zhang, and Markus Freitag. 2025. Wmt24++: Expanding the language coverage of wmt24 to 55 languages & dialects. ↗
  14. Duden. 2023. hudeln. Accessed: 2026-03-05. find
  15. Kenneth Enevoldsen, Isaac Chung, Imene Kerboua, Márton Kardos, Ashwin Mathur, David Stap, Jay Gala, Wissam Siblini, Dominik Krzemiński, Genta Indra Winata, et al. 2025. Mmteb: Massive multilingual text embedding benchmark. In The Thirteenth International Conference on Learning Representations. ↗
  16. Fangxiaoyu Feng, Yinfei Yang, Daniel Cer, Naveen Arivazhagan, and Wei Wang. 2022. Language-agnostic bert sentence embedding.
  17. Hermann Fischer. 1911. Schwäbisches Wörterbuch, volume 3. Verlag der Laupp’schen Buchhandlung, Tübingen. find
  18. Chetan Harsha, Karmvir Singh Phogat, Sridhar Dasaratha, Sai Akhil Puranam, and Shashishekar Ramakrishna. 2025. Synthetic data generation using large language models for financial question answering. In Proceedings of the Joint Workshop of the 9th Financial Technology and Natural Language Processing (FinNLP), the 6th Financial Narrative Processing (FNP), and the 1st Workshop on Large Language Models for Finance and Legal (LLMFinLegal), pages 76–95, Abu Dhabi, UAE. Association for Computational Linguistics. find
  19. Wan-hua Her and Udo Kruschwitz. 2024. Investigating neural machine translation for low-resource languages: Using Bavarian as a case study. In Proceedings of the 3rd Annual Meeting of the Special Interest Group on Under-resourced Languages @ LREC-COLING 2024, pages 155–167, Torino, Italia. ELRA and ICCL. find
  20. Vitor Jeronymo, Luiz Bonifacio, Hugo Abonizio, Marzieh Fadaee, Roberto Lotufo, Jakub Zavrel, and Rodrigo Nogueira. 2023. Inpars-v2: Large language models as efficient dataset generators for information retrieval. ↗
  21. Vladimir Karpukhin, Barlas Oguz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih. 2020. Dense passage retrieval for open-domain question answering. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 6769–6781, Online. Association for Computational Linguistics. ↗
  22. Seungone Kim, Juyoung Suk, Xiang Yue, Vijay Viswanathan, Seongyun Lee, Yizhong Wang, Kiril Gashteovski, Carolin Lawrence, Sean Welleck, and Graham Neubig. 2025. Evaluating language models as synthetic data generators. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 6385–6403, Vienna, Austria. Association for Computational Linguistics. ↗
  23. Julia Kreutzer, Isaac Caswell, Lisa Wang, Ahsan Wahab, Daan van Esch, Nasanbayar Ulzii-Orshikh, Allahsera Tapo, Nishant Subramani, Artem Sokolov, Claytone Sikasote, Monang Setyawan, Supheakmungkol Sarin, Sokhar Samb, Benoît Sagot, Clara Rivera, Annette Rios, Isabel Papadimitriou, Salomey Osei, Pedro Ortiz Suarez, Iroro Orife, Kelechi Ogueji, Andre Niyongabo Rubungo, Toan Q. Nguyen, Mathias Müller, André Müller, Shamsuddeen Hassan Muhammad, Nanda Muhammad, Ayanda Mnyakeni, Jamshidbek Mirzakhalov, Tapiwanashe Matangira, Colin Leong, Nze Lawson, Sneha Kudugunta, Yacine Jernite, Mathias Jenny, Orhan Firat, Bonaventure F. P. Dossou, Sakhile Dlamini, Nisansa de Silva, Sakine Çabuk Ballı, Stella Biderman, Alessia Battisti, Ahmed Baruwa, Ankur Bapna, Pallavi Baljekar, Israel Abebe Azime, Ayodele Awokoya, Duygu Ataman, Orevaoghene Ahia, Oghenefego Ahia, Sweta Agrawal, and Mofetoluwa Adeyemi. 2022. Quality at a glance: An audit of web-crawled multilingual datasets. Transactions of the Association for Computational Linguistics, 10:50– 72. ↗
  24. Niklas Laxström, Pau Giner, and Santhosh Thottingal. 2015. Content translation: Computer-assisted translation tool for wikipedia articles. In Proceedings of the 18th Annual Conference of the European Association for Machine Translation. ↗
  25. Seonghoon Lim, Taejun Yun, Jinhyeon Kim, Jihun Choi, and Taeuk Kim. 2024. Analysis of multi-source language training in cross-lingual transfer. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 712–725. find
  26. Robert Litschko, Verena Blaschke, Diana Burkhardt, Barbara Plank, and Diego Frassinelli. 2025a. Make every letter count: Building dialect variation dictionaries from monolingual corpora. find
  27. Robert Litschko, Goran Glavaš, Ivan Vulic, and Laura Dietz. 2019. Evaluating resource-lean cross-lingual embedding models in unsupervised retrieval. In Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR’19, page 1109–1112, New York, NY, USA. Association for Computing Machinery. find
  28. Robert Litschko, Oliver Kraus, Verena Blaschke, and Barbara Plank. 2025b. Cross-dialect information retrieval: Information access in low-resource and high-variance languages. ↗
  29. Ernst Martin and Hans Lienhart. 2012. Wörterbuch der elsässischen Mundarten. Walter de Gruyter. find
  30. Friedrich (ed.) Maurer, Friedrich Stroh, Rudolf Mulch, and Roland Mulch. 1973–1977. Südhessisches Wörterbuch, volume H–ksch of Hes-sische Historische Kommission Darmstadt. Marburg. find
  31. Josef Müller, Heinrich Dittmaier, Karl Meisen, and Matthias Zender. 1928–1971. Rheinisches wörterbuch, digitalisierte fassung im wörterbuchnetz des trier center for digital humanities. find
  32. Sreyashi Nag, Mihir Kale, Varun Lakshminarasimhan, and Swapnil Singhavi. 2020. Incorporating bilingual dictionaries for low resource semi-supervised neural machine translation. ↗
  33. Aaron van den Oord, Yazhe Li, and Oriol Vinyals. 2019. Representation learning with contrastive predictive coding. ↗
  34. OpenAI. 2024. Gpt-4o system card.
  35. Siyao Peng, Zihang Sun, Huangyan Shan, Marie Kolm, Verena Blaschke, Ekaterina Artemova, and Barbara Plank. 2024. Sebastian, Basti, Wastl?! recognizing named entities in Bavarian dialectal data. In Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024), pages 14478–14493, Torino, Italia. ELRA and ICCL. find
  36. Fred Philippy, Siwen Guo, Jacques Klein, and Tegawende Bissyande. 2025. LuxEmbedder: A cross-lingual approach to enhanced Luxembourgish sentence embeddings. In Proceedings of the 31st International Conference on Computational Linguistics, pages 11369–11379, Abu Dhabi, UAE. Association for Computational Linguistics. ↗
  37. Nils Reimers and Iryna Gurevych. 2019. Sentence-bert: Sentence embeddings using siamese bert-networks.
  38. Stephen Robertson and Hugo Zaragoza. 2009. The probabilistic relevance framework: BM25 and beyond, volume 4. Now Publishers Inc. find
  39. Barbara Scalvini, Iben Nyholm Debess, Annika Simonsen, and Hafsteinn Einarsson. 2025. Rethinking low-resource MT: the surprising effectiveness of fine-tuned multilingual models in the LLM age. In Proceedings of the Joint 25th Nordic Conference on Computational Linguistics and 11th Baltic Conference on Human Language Technologies (NoDaLiDa/Baltic-HLT 2025), pages 609–621, Tallinn, Estonia. University of Tartu Library. find
  40. Holger Schwenk, Vishrav Chaudhary, Shuo Sun, Hongyu Gong, and Francisco Guzmán. 2021. WikiMatrix: Mining 135M parallel sentences in 1620 language pairs from Wikipedia. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 1351–1361, Online. Association for Computational Linguistics.
  41. Nandan Thakur, Jianmo Ni, Gustavo Hernandez Abrego, John Wieting, Jimmy Lin, and Daniel Cer. 2024. Leveraging LLMs for synthesizing training data across many languages in multilingual dense retrieval. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pages 7699–7724, Mexico City, Mexico. Association for Computational Linguistics. ↗
  42. Jörg Tiedemann. 2012. Parallel data, tools and interfaces in opus. In Proceedings of the Eighth International Conference on Language Resources and Evaluation, volume 2012, pages 2214–2218. find
  43. Jörg Tiedemann. 2020. The tatoeba translation challenge – realistic data sets for low resource and multilingual MT. In Proceedings of the Fifth Conference on Machine Translation, pages 1174–1182, Online. Association for Computational Linguistics.
  44. Jannis Vamvas, Noëmi Aepli, and Rico Sennrich. 2024. Modular adaptation of multilingual encoders to written Swiss German dialect. In Proceedings of the 1st Workshop on Modular and Open Multilingual NLP (MOOMIN 2024), pages 16–23, St Julians, Malta. Association for Computational Linguistics. ↗
  45. Jannis Vamvas, Ignacio Pérez Prat, Not Battesta Soliva, Sandra Baltermia-Guetg, Andrina Beeli, Simona Beeli, Madlaina Capeder, Laura Decurtins, Gian Peder Gregori, Flavia Hobi, Gabriela Holderegger, Arina Lazzarini, Viviana Lazzarini, Walter Rosselli, Bettina Vital, Anna Rutkiewicz, and Rico Sennrich. 2025. Expanding the wmt24++ benchmark with rumantsch grischun, sursilvan, sutsilvan, surmiran, puter, and vallader. ↗
  46. Raul Vazquez, Timothee Mickus, Elaine Zosa, Teemu Vahtola, Jörg Tiedemann, Aman Sinha, Vincent Segonne, Fernando Sanchez Vega, Alessandro Raganato, Jindřich Libovický, Jussi Karlgren, Shaoxiong Ji, Jindřich Helcl, Liane Guillou, Ona De Gibert, Jaione Bengoetxea, Joseph Attieh, and Marianna Apidianaki. 2025. SemEval-2025 task 3: Mu-SHROOM, the multilingual shared-task on hallucinations and related observable overgeneration mistakes. In Proceedings of the 19th International Workshop on Semantic Evaluation (SemEval-2025), pages 2472–2497, Vienna, Austria. Association for Computational Linguistics. ↗
  47. Marcos Zampieri, Shervin Malmasi, Nikola Ljubešić, Preslav Nakov, Ahmed Ali, Jörg Tiedemann, Yves Scherrer, and Noëmi Aepli. 2017. Findings of the VarDial evaluation campaign 2017. In Proceedings of the Fourth Workshop on NLP for Similar Languages, Varieties and Dialects (VarDial), pages 1–15, Valencia, Spain. Association for Computational Linguistics.
  48. Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q. Weinberger, and Yoav Artzi. 2020. Bertscore: Evaluating text generation with bert. In International Conference on Learning Representations. ↗
  49. Xin Zhang, Yanzhao Zhang, Dingkun Long, Wen Xie, Ziqi Dai, Jialong Tang, Huan Lin, Baosong Yang, Pengjun Xie, Fei Huang, Meishan Zhang, Wenjie Li, and Min Zhang. 2024. mgte: Generalized long-context text representation and reranking models for multilingual text retrieval. ↗
  50. Yanzhao Zhang, Mingxin Li, Dingkun Long, Xin Zhang, Huan Lin, Baosong Yang, Pengjun Xie, An Yang, Dayiheng Liu, Junyang Lin, Fei Huang, and Jingren Zhou. 2025. Qwen3 embedding: Advancing text embedding and reranking through foundation models. ↗
  51. Shijia Zhou, Huangyan Shan, Barbara Plank, and Robert Litschko. 2024. Mainlp at semeval-2024 task 1: Analyzing source language selection in cross-lingual textual relatedness. In Proceedings of the 18th International Workshop on Semantic Evaluation (SemEval-2024), pages 1842–1853. find
  52. Pierre Zweigenbaum, Serge Sharoff, and Reinhard Rapp. 2017. Overview of the second BUCC shared task: Spotting parallel sentences in comparable corpora. In Proceedings of the Tenth Workshop on Building and Using Comparable Corpora, pages 60–67, Vancouver, Canada. Association for Computational Linguistics. find
  53. Pierre Zweigenbaum, Serge Sharoff, and Reinhard Rapp. 2018. Overview of the third bucc shared task: Spotting parallel sentences in comparable corpora. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018), Paris, France. European Language Resources Association (ELRA). find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/wang-2026-bi-text-mining-across-german/