Leveraging Digitized Newspapers to Collect Summarization Data in Low-Resource Languages

What it is about

Front-page teasers are short blurbs that a newspaper's editors write on the first page to point to full articles inside the issue. The authors show that teasers are common in seven languages: about nine issues per title were needed to collect 30 teaser-article pairs, and most titles have multi-document teasers. Expert annotators rate teasers highly on fluency and coherence, and relevance rises with length (3.29 for teasers under 25 words, 4.30 over 100). The collection is automated in two steps: a per-title list of key phrases finds the teasers (precision 95%, recall 85% on 102 pages), and the cited page's articles are matched to each teaser with TF-IDF and a threshold, a fine-tuned sentence transformer, or a zero-shot LLM. The zero-shot LLM matches best, with TF-IDF close behind and far cheaper. The result is HebTeaseSum: 7,774 Hebrew summaries from one 1984-1993 newspaper, 3,869 of them multi-document. The authors do not publish the texts, only the code and a way to rebuild them from the national library.

Ideas

• I1. Teaser detection as a finite list: a per-title set of key phrases and continuation words, applied to OCR text. Its errors (captions hiding the teaser, headers that copy the key word) show what the list must handle, and the method is deterministic.
• I2. TF-IDF with a threshold tuned on about 50 hand-labelled pairs as a cheap, transparent retrieval step for Hebrew. It is a baseline for matching a question to passages before any reasoning, and it needs no model.
• I3. Length as a visible quality signal: a summary under 25 words loses relevance, so a summary check can report length first and coverage second.

For Mova

I1 belongs to tools/ocrfix as a deterministic step on printed Hebrew. I2 is a baseline for the retrieval side of the RAG work, to be measured against a held-out set of Mova's own. I3 is a note for 12. Mova and Modern Hebrew: where it stands, what it would need if Mova later gets a summary task. The dataset itself is not to be carried into a release.

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-eacl.278

@inproceedings{dahan2026leveraging,
  title = {{Leveraging Digitized Newspapers to Collect Summarization Data in Low-Resource Languages}},
  author = {Dahan, Noam and Kidron, Omer and Stanovsky, Gabriel},
  booktitle = {Findings of the Association for Computational Linguistics: EACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-eacl.278},
  url = {https://doi.org/10.18653/v1/2026.findings-eacl.278}
}

References

47 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. AI@Meta. 2024. Llama 3 model card. find
2. Jean-Yves Antoine, Jeanne Villaneau, and Anaïs Lefeuvre. 2014. Weighted krippendorff’s alpha is a more reliable metrics for multi-coders ordinal annotations: experimental studies on emotion, opinion and coreference annotation. In EACL 2014, pages 10–p. find
3. Daniel Deutsch and Dan Roth. 2021. Understanding the extent to which content quality metrics measure the information quality of summaries. In Proceedings of the 25th Conference on Computational Natural Language Learning, pages 300–309. find
4. Melodee H Beals, Emily Bell, Ryan Cordell, Paul Fyfe, Isabel Galina Russell, Tessa Hauswedell, Clemens Neudecker, Julianne Nyhan, Mila Oiva, Sebastian Padó, et al. 2020. The atlas of digitised newspapers and metadata: Reports from oceanic exchanges. find
5. Maud Ehrmann, Matteo Romanello, Stefan Bircher, and Simon Clematide. 2020. Introducing the clef 2020 hipe shared task: Named entity recognition and linking on historical newspapers. In Advances in Information Retrieval: 42nd European Conference on IR Research, ECIR 2020, Lisbon, Portugal, April 14– 17, 2020, Proceedings, Part II 42, pages 524–532. Springer. find
6. Savita Bhat and Vasudeva Varma. 2023. Large language models as annotators: A preliminary evaluation for annotating low-resource language content. In Proceedings of the 4th Workshop on Evaluation and Comparison of NLP Systems, pages 100–107. find
7. Alexander R Fabbri, Wojciech Kryściński, Bryan Mc-Cann, Caiming Xiong, Richard Socher, and Dragomir Radev. 2021. Summeval: Re-evaluating summarization evaluation. Transactions of the Association for Computational Linguistics, 9:391–409. find
8. Nadav Borenstein, Karolina Stańczak, Thea Rolskov, Natália da Silva Perez, Natacha Klein Käfer, and Isabelle Augenstein. 2023. Measuring intersectional biases in historical documents. arXiv preprint arXiv:2305.12376. ↗
9. Jessica Forde, Ruochen Zhang, Lintang Sutawika, Alham Aji, Samuel Cahyawijaya, Genta Indra Winata, Minghao Wu, Carsten Eickhoff, Stella Biderman, and Ellie Pavlick. 2024. Re-evaluating evaluation for multilingual summarization. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 19476–19493. find
10. Emanuela Boros, Ahmed Hamdi, Elvys Linhares Pontes, Luis Adrián Cabrera-Diego, Jose G. Moreno, Nicolas Sidere, and Antoine Doucet. 2020. Alleviating digitization errors in named entity recognition for historical documents. In Proceedings of the 24th Conference on Computational Natural Language Learning, pages 431–441, Online. Association for Computational Linguistics. find
11. Gustavo Candela and Rafael C Carrasco. 2022. Discovering emerging topics in textual corpora of galleries, libraries, archives, and museums institutions. Journal of the Association for Information Science and Technology, 73(6):820–833. find
12. Jianpeng Cheng and Mirella Lapata. 2016. Neural summarization by extracting sentences and words. arXiv preprint arXiv:1603.07252. ↗
13. Jacob Cohen. 1960. A coefficient of agreement for nominal scales. Educational and psychological measurement, 20(1):37–46. ↗
14. Noam Dahan and Gabriel Stanovsky. 2025. The state and fate of summarization datasets: A survey. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pages 7259–7278, Albuquerque, New Mexico. Association for Computational Linguistics. find
15. Melissa Dell, Jacob Carlson, Tom Bryan, Emily Silcock, Abhishek Arora, Zejiang Shen, Luca D’Amico-Wong, Quan Le, Pablo Querubin, and Leander Heldring. 2023. American stories: A large-scale structured text dataset of historical us newspapers. Advances in Neural Information Processing Systems, 36:80744–80772. find
16. Shengxiang Gao, Fang Nan, Yongbing Zhang, Yuxin Huang, Kaiwen Tan, and Zhengtao Yu. 2025. A mixed-language multi-document news summarization dataset and a graphs-based extract-generate model. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pages 9255– 9265, Albuquerque, New Mexico. Association for Computational Linguistics. find
17. Demian Gholipour Ghalandari, Chris Hokamp, Nghia The Pham, John Glover, and Georgiana Ifrim. 2020. A large-scale multi-document summarization dataset from the wikipedia current events portal. Preprint, arXiv:2005.10070. ↗
18. Max Grusky, Mor Naaman, and Yoav Artzi. 2018. Newsroom: A dataset of 1.3 million summaries with diverse extractive strategies. arXiv preprint arXiv:1804.11283. ↗
19. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Ruoyu Zhang, Runxin Xu, Qihao Zhu, Shirong Ma, Peiyi Wang, Xiao Bi, et al. 2025. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948. ↗
20. Rishav Hada, Varun Gumma, Mohamed Ahmed, Kalika Bali, and Sunayana Sitaram. 2024. Metal: Towards multilingual meta-evaluation. arXiv preprint arXiv:2404.01667. ↗
21. Mohamed Haneefa and PT Jiji. 2019. Contents and interactivity of national library websites. DESI-DOC Journal of Library & Information Technology, 39(3):131. find
22. Karl Moritz Hermann, Tomas Kocisky, Edward Grefenstette, Lasse Espeholt, Will Kay, Mustafa Suleyman, and Phil Blunsom. 2015. Teaching machines to read and comprehend. Advances in neural information processing systems, 28. find
23. Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, et al. 2024. Gpt-4o system card. arXiv preprint arXiv:2410.21276. ↗
24. Albert Q Jiang, Alexandre Sablayrolles, Antoine Roux, Arthur Mensch, Blanche Savary, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Emma Bou Hanna, Florian Bressand, et al. 2024. Mixtral of experts. arXiv preprint arXiv:2401.04088. ↗
25. Pratik Joshi, Sebastin Santy, Amar Budhiraja, Kalika Bali, and Monojit Choudhury. 2020. The state and fate of linguistic diversity and inclusion in the nlp world. arXiv preprint arXiv:2004.09095. ↗
26. Kimmo Kettunen, Mika Koistinen, and Jukka Kervinen. 2020. Ground truth ocr sample data of finnish historical newspapers and journals in data improvement validation of a re-ocring process. LIBER Quarterly: The Journal of the Association of European Research Libraries, 30(1):1–20. find
27. Klaus Krippendorff. 2018. Content analysis: An introduction to its methodology. Sage publications. find
28. Haoyuan Li, Yusen Zhang, Rui Zhang, and Snigdha Chaturvedi. 2024. Coverage-based fairness in multi-document summarization. arXiv preprint arXiv:2412.08795. ↗
29. Chin-Yew Lin. 2004. Rouge: A package for automatic evaluation of summaries. In Text summarization branches out, pages 74–81. find
30. Yixin Liu, Alexander R Fabbri, Pengfei Liu, Yilun Zhao, Linyong Nan, Ruilin Han, Simeng Han, Shafiq Joty, Chien-Sheng Wu, Caiming Xiong, et al. 2022. Revisiting the gold standard: Grounding summarization evaluation with robust human evaluation. arXiv preprint arXiv:2212.07981. ↗
31. Itai Mondshine, Tzuf Paz-Argaman, Asaf Achi Mordechai, and Reut Tsarfaty. 2024. Hesum: a novel dataset for abstractive text summarization in hebrew. In Proceedings of the Seventh Workshop on Technologies for Machine Translation of Low-Resource Languages (LoResMT 2024), pages 26–36. find our card
32. Shashi Narayan, Shay B Cohen, and Mirella Lapata. 2018. Don’t give me the details, just the summary! topic-aware convolutional neural networks for extreme summarization. arXiv preprint arXiv:1808.08745. ↗
33. Xiao Pu, Mingqi Gao, and Xiaojun Wan. 2023. Summarization is (almost) dead. Preprint, arXiv:2309.09558. ↗
34. Nils Reimers and Iryna Gurevych. 2019. Sentence-bert: Sentence embeddings using siamese bert-networks. Preprint, arXiv:1908.10084. ↗
35. Chantal Shaib, Joe Barrow, Alexa F Siu, Byron C Wallace, and Ani Nenkova. 2024. How much annotation is needed to compare summarization models? arXiv preprint arXiv:2402.18756. ↗
36. Priyam Tejaswin, Dhruv Naik, and Pengfei Liu. 2021. How well do you know your summarization datasets? arXiv preprint arXiv:2106.11388. ↗
37. Alan Thomas, Robert Gaizauskas, and Haiping Lu. 2024. Leveraging llms for post-ocr correction of historical newspapers. In Proceedings of the Third Workshop on Language Technologies for Historical and Ancient Languages (LT4HALA)@ LREC-COLING-2024, pages 116–121. find
38. Reut Tsarfaty, Dan Bareket, Stav Klein, and Amit Seker. 2020. From spmrl to nmrl: What did we learn (and unlearn) in a decade of parsing morphologically-rich languages (mrls)? arXiv preprint arXiv:2005.01330. ↗
39. Sandra H Utt and Steve Pasternack. 1989. How they look: An updated study of american newspaper front pages. Journalism Quarterly, 66(3):621–627. find
40. Daniel Varab and Natalie Schluter. 2021. Massivesumm: a very large-scale, very multilingual, news summarisation dataset. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 10150–10161. find
41. Jiaan Wang, Yunlong Liang, Fandong Meng, Zengkui Sun, Haoxiang Shi, Zhixu Li, Jinan Xu, Jianfeng Qu, and Jie Zhou. 2023a. Is chatgpt a good nlg evaluator? a preliminary study. arXiv preprint arXiv:2303.04048. ↗
42. Jiaan Wang, Yunlong Liang, Fandong Meng, Beiqi Zou, Zhixu Li, Jianfeng Qu, and Jie Zhou. 2023b. Zero-shot cross-lingual summarization via large language models. arXiv preprint arXiv:2302.14229. ↗
43. Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, and Colin Raffel. 2021. mT5: A massively multilingual pre-trained text-to-text transformer. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 483–498, Online. Association for Computational Linguistics. find
44. Shiyue Zhang, Ben Frey, and Mohit Bansal. 2022. How can nlp help revitalize endangered languages? a case study and roadmap for the cherokee language. arXiv preprint arXiv:2204.11909. ↗
45. Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q Weinberger, and Yoav Artzi. 2019. Bertscore: Evaluating text generation with bert. arXiv preprint arXiv:1904.09675. ↗
46. Tianyi Zhang, Faisal Ladhak, Esin Durmus, Percy Liang, Kathleen McKeown, and Tatsunori B Hashimoto. 2024. Benchmarking large language models for news summarization. Transactions of the Association for Computational Linguistics, 12:39–57. find
47. Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric Xing, et al. 2023. Judging llm-as-a-judge with mt-bench and chatbot arena. Advances in Neural Information Processing Systems, 36:46595–46623. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/aclbulk-2026-dahan-leveraging-digitized-newspapers-collect-summarization/