Mova0.5 pre-release⌕
☰
home / docs / papers / Ancient Hebrew / Systematic Textual Availability of Manuscripts

Systematic Textual Availability of Manuscripts

What it is about

Hebrew manuscripts are many, but handwritten text recognition still does not turn them into usable text at scale. The authors build a pipeline: a pretrained kraken-type recognition model is fine-tuned on the first few labelled pages of each manuscript, and the output is then corrected in two ways. The first is text-reuse detection against a reference library (Sefaria): a fuzzy search on an inverted index, approximate alignment for noisy OCR, and a majority vote over the aligned readings. The second is spell-checking and rescoring with BEREL, a BERT model for Rabbinic Hebrew, and with a generative model. On a 14th-century Midrash manuscript, fine-tuning raises character accuracy by about 18% (as the abstract states), and in a preliminary run text reuse lifts word accuracy from about 81% to 91%. Rescoring with generative models was unstable, while text reuse combined with spell-checking did best.

Ideas

For Mova

I1 and I2 belong in tools/ocrfix (correction of OCR of printed or handwritten Hebrew and Aramaic, with deterministic rules), and I1 can also feed the unpointed mode of tools/hbo on raw text. I3 is a measurement for tools/ocrfix. The paper's finding that the generative rescoring is unstable supports keeping language models out of the deterministic core.

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.5281/zenodo.18051490

@inproceedings{miller2025systematic,
  title = {{Systematic Textual Availability of Manuscripts}},
  author = {Miller, Hadar and Londner, Samuel and Kuflik, Tsvi and Shapira, Daria Vasyutinsky and Dershowitz, Nachum and Lavee, Moshe},
  booktitle = {Proceedings of the 5th Conference on Language, Data and Knowledge},
  year = {2025},
  doi = {10.5281/zenodo.18051490},
  url = {https://doi.org/10.5281/zenodo.18051490}
}

References

68 references, 9 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Wissam AlKendi, Franck Gechter, Laurent Heyberger, and Christophe Guyeux. 2024. Advancements and challenges in handwritten text recognition: A comprehensive survey. Journal of Imaging, 10(1):18. find
  2. Jonathan Parkes Allen, Matthew Thomas Miller, John Mullan, and David Smith. 2022. Digitizing the Islamicate written traditions: History, state of the field, and best practices for open-source Arabic-script OCR. White paper AOCP Phase I White Paper v. 1.1, Open-ITI. find
  3. Stephen F. Altschul, Warren Gish, Webb Miller, Eugene W. Myers, and David J. Lipman. 1990. Basic local alignment search tool. Journal of Molecular Biology, 215(3):403–410. find
  4. Adam Amram, Anat Ben David, and Reut Tsarfaty. 2018. Representations and architectures in neural sentiment analysis for morphologically rich languages: A case study from Modern Hebrew. In Proc. of the 27th International Conference on Computational Linguistics, pages 2242–2252. find our card
  5. José Carlos Aradillas, Juan José Murillo-Fuentes, and Pablo M. Olmos. 2021. Boosting offline handwritten text recognition in historical documents with few labeled lines. IEEE Access, 9:76674–76688. find
  6. Adiel Ben-Shalom, Adi Silberpfennig, Nachum Dershowitz, Lior Wolf, and Yaacov Choueka. 2017. Querying Hebrew texts via word spotting. In World Congress of Jewish Studies, Jerusalem, Israel. Biblioteca Apostolica Vaticana. Midrash Tanḥuma. Ms. Vat.ebr.44.pt.1. Bibliothèque de Genève. Midrash Tanhuma (Leviticus-Numbers-Deuteronomy). Ms. Comites Latentes 146. find
  7. Oran Brill, Moshe Koppel, and Avi Shmidman. 2020. FAST: Fast and accurate synoptic texts. Digital Scholarship in the Humanities, 35(2):254–264. find
  8. Marco Büchler, Philip R. Burns, Martin Müller, Emily Franzini, and Greta Franzini. 2014. Towards a historical text re-use detection. In Text Mining, pages 221–238. Springer. find
  9. Avihay Chriqui and Inbal Yahav. 2022. HeBERT and HebEMO: A Hebrew BERT model and a tool for polarity analysis and emotion recognition. INFORMS Journal on Data Science, 1(1):81–95. ↗ our card
  10. Paul Clough, Robert Gaizauskas, Scott S. L. Piao, and Yorick Wilks. 2002. METER: MEasuring TExt Reuse. In Proceedings of 40th Anniversary Meeting for the Association for Computational Linguistics, pages 152–159. find
  11. Rafi Cohen, Irina Rabaev, Jihad El-Sana, Klara Kedem, and Itshak Dinstein. 2015. Aligning transcript of historical documents using energy minimization. In Proc. 13th International Conference on Document Analysis and Recognition (ICDAR), pages 266–270. IEEE. find
  12. Giovanni Colavizza, Mario Infelise, and Frédéric Kaplan. 2014. Mapping the early modern news flow: An enquiry by robust text reuse detection. In International Conference on Social Informatics, pages 244–253. Springer. find
  13. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. Bert: Pre-training of deep bidirectional transformers for language understanding. In Proc. of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, volume 1, pages 4171–4186. find
  14. Kartik Dutta, Praveen Krishnan, Minesh Mathew, and C. V. Jawahar. 2018. Improving CNN-RNN hybrid networks for handwriting recognition. In 16th International Conference on Frontiers in Handwriting Recognition, pages 80–85. IEEE. find
  15. Tomáš Foltỳnek, Norman Meuschke, and Bela Gipp. 2019. Academic plagiarism detection: A systematic literature review. ACM Computing Surveys, 52(6):1–42. find
  16. Shaona Ghosh and Per Ola Kristensson. 2017. Neural networks for text correction and completion in keyboard decoding. arXiv preprint arXiv:1709.06429. ↗
  17. Ian Goodfellow, Yoshua Bengio, and Aaron Courville. 2016. Deep Learning. MIT Press. find
  18. Adeline Granet, Emmanuel Morin, Harold Mouchère, Solen Quiniou, and Christian Viard-Gaudin. 2018. Transfer learning for handwriting recognition on historical documents. In 7th International Conference on Pattern Recognition Applications and Methods (ICPRAM). find
  19. Eylon Gueta, Avi Shmidman, Shaltiel Shmidman, Cheyn Shmuel Shmidman, Joshua Guedalia, Moshe Koppel, Dan Bareket, Amit Seker, and Reut Tsarfaty. 2022. Large pre-trained models with extra-large vocabularies: A contrastive analysis of Hebrew BERT models and a new one to outperform them all. arXiv preprint arXiv:2211.15199. ↗ our card
  20. Yaakov HaCohen-Kerner, Ariel Kass, and Ariel Peretz. 2004. Baseline methods for automatic disambiguation of abbreviations in Jewish law documents. In Advances in Natural Language Processing, pages 58–69. Springer. find
  21. Yaakov HaCohen-Kerner, Nadav Schweitzer, and Dror Mughaz. 2011. Automatically identifying citations in Hebrew-Aramaic documents. Cybernetics and Systems: An International Journal, 42(3):180–197. find
  22. Kai Hakala, Aleksi Vesanto, Niko Miekka, Tapio Salakoski, and Filip Ginter. 2019. Leveraging text repetitions and denoising autoencoders in OCR post-correction. arXiv preprint arXiv:1906.10907. ↗
  23. John C. Handley and Thomas B. Hickey. 1991. Merging optical character recognition outputs for improved accuracy. In RIAO ’91: Intelligent Text and Image Handling, pages 160–174, Paris. Le Centre de Hautes Etudes Internationales d’Informatique Documentaire. find
  24. Awni Hannun. 2017. Sequence modeling with CTC. Distill. https://distill.pub/2017/ctc. ↗
  25. Awni Y. Hannun, Andrew L. Maas, Daniel Jurafsky, and Andrew Y. Ng. 2014. First-pass large vocabulary continuous speech recognition using bi-directional recurrent DNNs. arXiv preprint arXiv:1408.2873. ↗
  26. Alon Itai and Shuly Wintner. 2008. Language resources for Hebrew. Language Resources and Evaluation, 42(1):75–98. find
  27. José Carlos Aradillas Jaramillo, Juan José Murillo-Fuentes, and Pablo M. Olmos. 2018. Boosting handwriting text recognition in small databases with transfer learning. In Proc. 16th International Conference on Frontiers in Handwriting Recognition (ICFHR), pages 429–434. IEEE. find
  28. Philip Kahle, Sebastian Colutto, Günter Hackl, and Günter Mühlberger. 2017. Transkribus – A service platform for transcription, recognition and retrieval of historical documents. In Proc. 14th IAPR International Conference on Document Analysis and Recognition (ICDAR), volume 4, pages 19–24. IEEE. find
  29. Omri Keren, Tal Avinari, Reut Tsarfaty, and Omer Levy. 2022. Breaking character: Are subwords good enough for MRLs after all? arXiv preprint arXiv:2204.04748. ↗
  30. Benjamin Kiessling. 2019. Kraken – an universal text recognizer for the humanities. In Digital Humanities (DH2019). find
  31. Benjamin Kiessling, Robin Tissot, Peter Stokes, and Daniel Stökl Ben Ezra. 2019. eScriptorium: An open source platform for historical document analysis. In Proc. International Conference on Document Analysis and Recognition Workshops (ICDARW), volume 2, pages 19–24. IEEE. find
  32. Ido Kissos and Nachum Dershowitz. 2016. OCR error correction using character correction and feature-based word classification. In Proc. 12th IAPR Workshop on Document Analysis Systems (DAS), pages 198–203. IEEE. find
  33. Ido Kissos and Nachum Dershowitz. 2017. Image and text correction using language models. In Proc. 1st International Workshop on Arabic Script Analysis and Recognition, pages 158–162. IEEE. find
  34. Benjamin Eliot Klein, Nachum Dershowitz, Lior Wolf, Orna Almogi, and Dorji Wangchuk. 2014. Finding inexact quotations within a Tibetan Buddhist corpus. In Digital Humanities, pages 486–488. find
  35. Tsvi Kuflik, Moshe Lavee, Daniel Stoekl Ben Ezra, Avigail Ohali, Vered Raziel-Kretzmer, Uri Schor, Alan Wecker, Elena Lolli, and Pauline Signoret. 2019. Tikkoun Sofrim combining HTR and crowdsourcing for automated transcription of Hebrew medieval manuscripts. In Digital Humanities (DH2019). find
  36. Karen Kukich. 1992. Techniques for automatically correcting words in text. ACM Computing Surveys (CSUR), 24(4):377–439. find
  37. Berat Kurar Barakat, Jihad El-Sana, and Irina Rabaev. 2019. The Pinkas dataset. In Proc. International Conference on Document Analysis and Recognition (ICDAR), pages 732–737. IEEE. find
  38. Sarah Bowen Savant. 2016. The history of Arabic books in the digital age. British Academy Review, 28. find
  39. Jey Han Lau, Carlos Armendariz, Shalom Lappin, Matthew Purver, and Chang Shu. 2020. How furiously can colorless green ideas sleep? Sentence acceptability in context. Transactions of the Association for Computational Linguistics, 8:296–310. find
  40. Amit Seker, Elron Bandel, Dan Bareket, Idan Brusilovsky, Refael Greenfeld, and Reut Tsarfaty. 2022. AlephBERT: Language model pre-training and evaluation from sub-word to sentence level. In Proc. of the 60th Annual Meeting of the Association for Computational Linguistics, volume 1, pages 46–56. find our card
  41. John S. Y. Lee. 2007. A computational model of text reuse in ancient literary texts. In Proc. of the 45th Annual Meeting of the Association of Computational Linguistics, pages 472–479. find
  42. Vladimir I. Levenshtein. 1966. Binary codes capable of correcting deletions, insertions, and reversals. Soviet Physics Doklady, 10(8):707–710. In Russian, translated into English. find
  43. Shahar Mahpod and Yosi Keller. 2018. Auto-ML deep learning for Rashi scripts OCR. CoRR, abs/1811.01290. ↗
  44. Hadar Miller, Tsvi Kuflik, and Moshe Lavee. 2025. Text alignment in the service of text reuse detection. Applied Sciences, 15(6). find
  45. Amir More, Amit Seker, Victoria Basmova, and Reut Tsarfaty. 2019. Joint transition-based models for morpho-syntactic parsing: Parsing strategies for MRLs and a case study from Modern Hebrew. Transactions of the Association for Computational Linguistics, 7:33–48. find our card
  46. National Library of Israel. 2021. Digitized Hebrew manuscripts. find
  47. Thi Tuyet Hai Nguyen, Adam Jatowt, Nhu-Van Nguyen, Mickael Coustaty, and Antoine Doucet. 2020. Neural machine translation with BERT for post-OCR error detection and correction. In Proc. of the ACM/IEEE Joint Conference on Digital Libraries, pages 333–336. find
  48. Franz Josef Och and Hermann Ney. 2003. A systematic comparison of various statistical alignment models. Computational Linguistics, 29(1):19–51. find
  49. Christian Reul, Dennis Christ, Alexander Hartelt, Nico Balbach, Maximilian Wehner, Uwe Springmann, Christoph Wick, Christine Grundig, Andreas Büttner, and Frank Puppe. 2019. OCR4all—An open-source tool providing a (semi-)automatic OCR workflow for historical printings. Applied Sciences, 9(22). find
  50. Sefaria, Inc. 2021. A living library of Torah texts online. find
  51. Avi Shmidman, Joshua Guedalia, Shaltiel Shmidman, Cheyn Shmuel Shmidman, Eli Handel, and Moshe Koppel. 2022. Introducing BEREL: BERT embeddings for rabbinic-encoded language. arXiv preprint arXiv:2208.01875. ↗ our card
  52. Shaltiel Shmidman, Avi Shmidman, Amir D. N. Cohen, and Moshe Koppel. 2024. Adapting LLMs to Hebrew: Unveiling DictaLM 2.0 with enhanced vocabulary and instruction capabilities. arXiv preprint arXiv:2407.07080. ↗ our card
  53. Shaltiel Shmidman, Avi Shmidman, Amir David Nissan Cohen, and Moshe Koppel. 2023. Introducing DictaLM – A large generative language model for Modern Hebrew. Preprint, arXiv:2309.14568. ↗ our card
  54. Michal Bar-Asher Siegal and Avi Shmidman. 2018. Reconstruction of the Mekhilta Deuteronomy using philological and computational tools. Journal of Ancient Judaism, 9(1):2–25. find
  55. David A. Smith, Ryan Cordel, Elizabeth Maddock Dillon, Nick Stramp, and John Wilkerson. 2014. Detecting and modeling local text reuse. In IEEE/ACM Joint Conference on Digital Libraries, pages 183–192. find
  56. David A. Smith, Ryan Cordell, and Elizabeth Maddock Dillon. 2013. Infectious texts: Modeling text reuse in nineteenth-century newspapers. In IEEE International Conference on Big Data, pages 86–94. find
  57. David A. Smith, Ryan Cordell, and Abby Mullen. 2015. Computational methods for uncovering reprinted texts in antebellum newspapers. American Literary History, 27(3):E1–E15. find
  58. Daniel Stökl Ben Ezra, Bronson Brown-DeVost, Pawel Jablonski, Hayim Lapin, Benjamin Kiessling, and Elena Lolli. 2021. BiblIA – a general model for medieval Hebrew manuscripts and an open annotated dataset. In The 6th International Workshop on Historical Document Imaging and Processing, pages 61–66. find
  59. Stephen V. Rice, Junichi Kanai, and Thomas A. Nartker. 1994. An algorithm for matching OCR-generated text strings. In Document Image Analysis, pages 263–272. World Scientific. find
  60. Omri Suissa, Avshalom Elmalech, and Maayan Zhitomirsky-Geffet. 2020. Optimizing the neural network training for OCR error correction of historical Hebrew texts. In iConference 2020 Proceedings. iSchools. find our card
  61. Julian Salazar, Davis Liang, Toan Q. Nguyen, and Katrin Kirchhoff. 2020. Masked language model scoring. In Proc. of the 58th Annual Meeting of the Association for Computational Linguistics. find
  62. Daria Vasyutinsky-Shapira, Berat Kurar-Barakat, Sharva Gogawale, Mohammad Suliman, and Nachum Dershowitz. 2024. MiDRASH – A project for computational analysis of medieval Hebrew manuscripts. In EUROGRAPHICS Workshop on Graphics and Cultural Heritage. find
  63. Aleksi Vesanto, Asko Nivala, Heli Rantala, Tapio Salakoski, Hannu Salmi, and Filip Ginter. 2017. Applying BLAST to text reuse detection in Finnish newspapers and journals, 1771–1910. In Proceedings of the NoDaLiDa 2017 Workshop on Processing Historical Language, pages 54–58, Gothenburg. LiU Electronic Press. find
  64. Mauricio Villegas, Alejandro H. Toselli, Verónica Romero, and Enrique Vidal. 2016. Exploiting existing modern transcripts for historical handwritten text recognition. In Proc. 15th International Conference on Frontiers in Handwriting Recognition (ICFHR), pages 66–71. IEEE. find
  65. Alan J. Wecker, Vered Raziel-Kertzmer, Daniel Stökl Ben Ezra, Moshe Lavee, Tsvi Kuflik, Dror Elovits, Moshe Schorr, Uri Schor, and Pawel Jablonski. 2022. Tikkoun Sofrim: Making ancient manuscripts digitally accessible: The case of Midrash Tanhuma. ACM Journal of Computation and Cultural Heritage, 15(2). find
  66. John Wilkerson, David Smith, and Nicholas Stramp. 2015. Tracing the flow of policy ideas in legislatures: A text reuse approach. American Journal of Political Science, 59(4):943–956. find
  67. Thomas Zenkel, Ramon Sanabria, Florian Metze, Jan Niehues, Matthias Sperber, Sebastian Stüker, and Alex Waibel. 2017. Comparison of decoding strategies for CTC acoustic models. arXiv preprint arXiv:1708.04469. ↗
  68. Alex Zhicharevich. 2011. Tools to aid OCR of Hebrew character manuscripts. Master’s thesis, The Blavatnik School of Computer Science, Tel Aviv University, February. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/aclbulk-2025-miller-systematic-textual-availability-manuscripts/