[Mova](https://md.mova.elamur.ai/free/en/) · [this page in HTML](https://mova.elamur.ai/free/en/docs/papers/hbo-2023-suissa-optimizing-neural-network-training-ocr/)

[home](https://md.mova.elamur.ai/free/en/) / [docs](https://md.mova.elamur.ai/free/en/docs/) / [papers](https://md.mova.elamur.ai/free/en/docs/papers/) / [Ancient Hebrew](https://md.mova.elamur.ai/free/en/docs/papers/hbo/) / Optimizing the Neural Network Training for OCR Error Correction of Historical Hebrew Texts

# Optimizing the Neural Network Training for OCR Error Correction of Historical Hebrew Texts

**Authors:** [Omri Suissa](https://md.mova.elamur.ai/free/en/docs/papers/authors/omri-suissa/), [Avshalom Elmalech](https://md.mova.elamur.ai/free/en/docs/papers/authors/avshalom-elmalech/), [Maayan Zhitomirsky-Geffet](https://md.mova.elamur.ai/free/en/docs/papers/authors/maayan-zhitomirsky-geffet/) · **Year:** 2023 · **Venue:** arXiv 2307.16220 (iConference-style short paper) **Link:** [arxiv.org/abs/2307.16220](https://arxiv.org/abs/2307.16220) · **Licence:** [CC BY-NC-ND 4.0](https://creativecommons.org/licenses/by-nc-nd/4.0/) — summary only, a measure at most

[**2** citations](https://openalex.org/works?filter=cites:W4385473920) openalex · 2026-10-18[**1** citations in our library](https://md.mova.elamur.ai/free/en/docs/papers/cited-by/hbo-2023-suissa-optimizing-neural-network-training-ocr/) of the 140673 papers whose bibliographies we read[arXiv:2307.16220](https://openalex.org/S4306400194)[licence](http://creativecommons.org/licenses/by-nc-nd/4.0/)

- [Publisher's page](https://arxiv.org/abs/2307.16220)
- [PDF](https://arxiv.org/pdf/2307.16220)
- [arXiv 2307.16220](https://arxiv.org/abs/2307.16220)
- [OpenAlex](https://openalex.org/W4385473920)

**Authors (3)**

1. [Omri Suissa](https://openalex.org/A5081395585) [ORCID](https://orcid.org/0000-0002-9703-2222) · h 4 · 109 citations · 11 works
2. [Avshalom Elmalech](https://openalex.org/A5047276842) [ORCID](https://orcid.org/0000-0001-6142-3881) · h 9 · 288 citations · 56 works
3. Maayan Zhitomirsky-Geffet

Citations and author metrics from openalex (2026-10-18); authors with the same name are not always told apart.

## What it is about

OCR of historical Hebrew newspapers (the JPress collection) is noisy, and neural post-correctors need large hand-corrected sets. The authors learn the typical OCR errors of the corpus from about a hundred hand-corrected articles, inject those errors into clean text to make large artificial training pairs, and train a light neural corrector. Training on corpus-specific injected errors beats random noise, and the result depends strongly on the period and genre of the clean text: trained on the Bible, the corrector introduces more errors than it fixes on modern-era newspapers. It also beats general spellcheckers on this task. This is the short version of the period-specific study (hbo-2023-suissa-toward-period-specific-optimized-neural).

## Ideas

- I1. Period matters for any Hebrew lexicon-based correction: a word list from the Torah must not be used to correct or normalise later Hebrew, nor the reverse. A rule for Mova's own lookups: each lexicon carries its period, and a lookup outside the period is flagged.
- I2. Errors learned from a small hand-corrected sample, then injected at scale, is a cheap way to build test sets; Mova can do the same for its damaged-text robustness test (see the companion card's confusion table).

## For Mova

Mova does no OCR; the neural corrector is out of scope and the NC-ND licence keeps the paper to tests and citation. I1 is a note in [12. Biblical, Mishnaic and Modern Hebrew compared](https://md.mova.elamur.ai/free/en/docs/languages/hbo/12-biblical-mishnaic-modern/) (why tools for one stage misread another) and a field in `macula dict`. I2 is the method for a test harness only.

## Cite

The paper itself, not this card. BibTeX from the library's record of the work. [arXiv 2307.16220](https://arxiv.org/abs/2307.16220)

```
@misc{suissa2023optimizing,
  title = {{Optimizing the Neural Network Training for OCR Error Correction of Historical Hebrew Texts}},
  author = {Suissa, Omri and Elmalech, Avshalom and Zhitomirsky-Geffet, Maayan},
  year = {2023},
  eprint = {2307.16220},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2307.16220}
}
```

- [← previous "You are Cursed by the God YHW": an early Hebrew inscription from Mt. Ebal](https://md.mova.elamur.ai/free/en/docs/papers/hbo-2023-stripling-cursed-god-yhw-early-hebrew/)
- [next → Toward a Period-Specific Optimized Neural Network for OCR Error Correction of Historical Hebrew Texts](https://md.mova.elamur.ai/free/en/docs/papers/hbo-2023-suissa-toward-period-specific-optimized-neural/)
- [↑ back to Ancient Hebrew](https://md.mova.elamur.ai/free/en/docs/papers/hbo/)

## References

20 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Chiron, G., Doucet, A., & Moreux, J.: Competition on Post-OCR Text Correction. In: ICDAR2017 Competition on Post-OCR Text Correction, pp. 1423–1428. (2017). [find](https://scholar.google.com/scholar?q=Chiron%2C+G%2E%2C+Doucet%2C+A%2E%2C+%26+Moreux%2C+J%2E%3A+Competition+on+Post%2DOCR+Text+Correction%2E+In%3A+ICDAR2017+Competition+on+Post%2DOCR+Text+Correction%2C+pp%2E+1423%E2%80%931428%2E+%282017%29%2E)
2. Cho, K., van Merrienboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., & Bengio, Y.: Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation. In: Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) (2014). [find](https://scholar.google.com/scholar?q=Cho%2C+K%2E%2C+van+Merrienboer%2C+B%2E%2C+Gulcehre%2C+C%2E%2C+Bahdanau%2C+D%2E%2C+Bougares%2C+F%2E%2C+Schwenk%2C+H%2E%2C+%26+Bengio%2C+Y%2E%3A+Learning+Phrase+Representations+using+RNN+Encoder%2DDecoder+for+Statistical+Machine+Translation%2E+In%3A+Pr)
3. Ghosh, S., & Kristensson, P. O.: Neural Networks for Text Correction and Completion in Keyboard Decoding. (2017). [find](https://scholar.google.com/scholar?q=Ghosh%2C+S%2E%2C+%26+Kristensson%2C+P%2E+O%2E%3A+Neural+Networks+for+Text+Correction+and+Completion+in+Keyboard+Decoding%2E+%282017%29%2E)
4. D’hondt, E., Grouin, C., & Grau, B.: Generating a Training Corpus for OCR Post-Correction Using Encoder-Decoder Model. In: Proceedings of the Eighth International Joint Conference on Natural Language Processing. vol. 1, pp. 1006-1014. (2017). [find](https://scholar.google.com/scholar?q=D%E2%80%99hondt%2C+E%2E%2C+Grouin%2C+C%2E%2C+%26+Grau%2C+B%2E%3A+Generating+a+Training+Corpus+for+OCR+Post%2DCorrection+Using+Encoder%2DDecoder+Model%2E+In%3A+Proceedings+of+the+Eighth+International+Joint+Conference+on+Natural+Language+)
5. Hochreiter, S., & Schmidhuber, J. J.: Long short-term memory. Neural Computation, 9(8), 1–32 (1997). [find](https://scholar.google.com/scholar?q=Hochreiter%2C+S%2E%2C+%26+Schmidhuber%2C+J%2E+J%2E%3A+Long+short%2Dterm+memory%2E+Neural+Computation%2C+9%288%29%2C+1%E2%80%9332+%281997%29%2E)
6. Lansdall-Welfare, T., Sudhahar, S., Thompson, J., Lewis, J., & Cristianini, N.: Content analysis of 150 years of British periodicals. In: Proceedings of the National Academy of Sciences. vol. 114 (4), pp. 457–465. (2017) [find](https://scholar.google.com/scholar?q=Lansdall%2DWelfare%2C+T%2E%2C+Sudhahar%2C+S%2E%2C+Thompson%2C+J%2E%2C+Lewis%2C+J%2E%2C+%26+Cristianini%2C+N%2E%3A+Content+analysis+of+150+years+of+British+periodicals%2E+In%3A+Proceedings+of+the+National+Academy+of+Sciences%2E+vol%2E+114+%284%29%2C)
7. Levenshtein, Vladimir I.: Binary codes capable of correcting deletions, insertions, and reversals. Soviet physics doklady 10(8), 707-710 (1966). [find](https://scholar.google.com/scholar?q=Levenshtein%2C+Vladimir+I%2E%3A+Binary+codes+capable+of+correcting+deletions%2C+insertions%2C+and+reversals%2E+Soviet+physics+doklady+10%288%29%2C+707%2D710+%281966%29%2E)
8. Needle, S. B., Christus, A. S. D., & Needleman, Saul B., and C. D. W.: A General Method Applicable to the Search for Similarities in the Amino Acid Sequence of Two Proteins. Journal of Molecular Biology, 48(3), 443–453 (1970). [find](https://scholar.google.com/scholar?q=Needle%2C+S%2E+B%2E%2C+Christus%2C+A%2E+S%2E+D%2E%2C+%26+Needleman%2C+Saul+B%2E%2C+and+C%2E+D%2E+W%2E%3A+A+General+Method+Applicable+to+the+Search+for+Similarities+in+the+Amino+Acid+Sequence+of+Two+Proteins%2E+Journal+of+Molecular+Biolo)
9. Pascanu, R., Mikolov, T., & Bengio, Y.: On the difficulty of training Recurrent Neural Networks. In: International conference on machine learning, pp. 1310-1318. (2013). [find](https://scholar.google.com/scholar?q=Pascanu%2C+R%2E%2C+Mikolov%2C+T%2E%2C+%26+Bengio%2C+Y%2E%3A+On+the+difficulty+of+training+Recurrent+Neural+Networks%2E+In%3A+International+conference+on+machine+learning%2C+pp%2E+1310%2D1318%2E+%282013%29%2E)
10. Raaijmakers, S.: A Deep Graphical Model for Spelling Correction. In: BNAIC 2013, Proceedings of the 25th Benelux Conference on Artificial Intelligence. Delft, The Netherlands (2013). [find](https://scholar.google.com/scholar?q=Raaijmakers%2C+S%2E%3A+A+Deep+Graphical+Model+for+Spelling+Correction%2E+In%3A+BNAIC+2013%2C+Proceedings+of+the+25th+Benelux+Conference+on+Artificial+Intelligence%2E+Delft%2C+The+Netherlands+%282013%29%2E)
11. Reynaert, M.: Non-interactive OCR post-correction for giga-scale digitization projects. In: Lecture Notes in Computer Science, LNCS, vol. 4919, pp. 617–630. Springer, Heidelberg (2008). [find](https://scholar.google.com/scholar?q=Reynaert%2C+M%2E%3A+Non%2Dinteractive+OCR+post%2Dcorrection+for+giga%2Dscale+digitization+projects%2E+In%3A+Lecture+Notes+in+Computer+Science%2C+LNCS%2C+vol%2E+4919%2C+pp%2E+617%E2%80%93630%2E+Springer%2C+Heidelberg+%282008%29%2E)
12. Schuster, M., & Paliwal, K. K.: Bidirectional recurrent neural networks. IEEE Transactions on Signal Processing, 45(11), 2673–2681 (1997). [find](https://scholar.google.com/scholar?q=Schuster%2C+M%2E%2C+%26+Paliwal%2C+K%2E+K%2E%3A+Bidirectional+recurrent+neural+networks%2E+IEEE+Transactions+on+Signal+Processing%2C+45%2811%29%2C+2673%E2%80%932681+%281997%29%2E)
13. Sutskever, I., Vinyals, O., & Le, Q. V.: Sequence to Sequence Learning with Neural Networks. In Advances in neural information processing systems, 3104-3112 (2014) [find](https://scholar.google.com/scholar?q=Sutskever%2C+I%2E%2C+Vinyals%2C+O%2E%2C+%26+Le%2C+Q%2E+V%2E%3A+Sequence+to+Sequence+Learning+with+Neural+Networks%2E+In+Advances+in+neural+information+processing+systems%2C+3104%2D3112+%282014%29)
14. Werbos, P.: Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Sciences. Ph. D. dissertation, Harvard University. (1974). [find](https://scholar.google.com/scholar?q=Werbos%2C+P%2E%3A+Beyond+Regression%3A+New+Tools+for+Prediction+and+Analysis+in+the+Behavioral+Sciences%2E+Ph%2E+D%2E+dissertation%2C+Harvard+University%2E+%281974%29%2E)
15. The British Newspaper Archive, [https://www.britishnewspaperarchive.co.uk/](https://www.britishnewspaperarchive.co.uk/), last accessed 2019/06/20. [↗](https://www.britishnewspaperarchive.co.uk/)
16. NYTimes.com, "Google Books: A Complex and Controversial Experiment", last accessed 2015/10/28. [find](https://scholar.google.com/scholar?q=NYTimes%2Ecom%2C+%22Google+Books%3A+A+Complex+and+Controversial+Experiment%22%2C+last+accessed+2015%2F10%2F28%2E)
17. JPress Collection Homepage, [http://web.nli.org.il/sites/JPress/Hebrew/Pages/default.aspx](http://web.nli.org.il/sites/JPress/Hebrew/Pages/default.aspx), last accessed 2019/06/20. [↗](http://web.nli.org.il/sites/JPress/Hebrew/Pages/default.aspx)
18. Project Ben Yehuda Homepage, [https://bybe.benyehuda.org/](https://bybe.benyehuda.org/), last accessed 2019/06/20. [↗](https://bybe.benyehuda.org/)
19. Project Gutenberg Homepage, [https://www.gutenberg.org/](https://www.gutenberg.org/), last accessed 2019/06/20. [↗](https://www.gutenberg.org/)
20. Ali, A., & Renals, S. (2018, July). Word error rate estimation for speech recognition: e-WER. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) (pp. 20-24 [find](https://scholar.google.com/scholar?q=Ali%2C+A%2E%2C+%26+Renals%2C+S%2E+%282018%2C+July%29%2E+Word+error+rate+estimation+for+speech+recognition%3A+e%2DWER%2E+In+Proceedings+of+the+56th+Annual+Meeting+of+the+Association+for+Computational+Linguistics+%28Volume+2%3A+Shor)