---
title: "From Tablet to Text — paper by Mahmood and Tetiana (2026)"
description: "The authors therefore present AI as a supervised junior collaborator, not an authority. 14 references."
authors: ["Jasim Mohammed Mahmood", "Panok Tetiana"]
year: 2026
venue: "Iraqi Literary and Cultural Review 3"
source: "https://ilcr.ibirc-center.co.uk/index.php/journal/article/view/30"
doi: "10.69513/ilcr.v3.i2.a4"
licence: "CC BY (journal licence per DOAJ; Iraqi Literary and Cultural Review)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/mahmood-2026-tablet-text-ai-assisted-reconstruction/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/mahmood-2026-tablet-text-ai-assisted-reconstruction/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/34/) / From Tablet to Text: AI-Assisted Reconstruction and Translation of Akkadian Passages in the Epic of Gilgamesh

# From Tablet to Text: AI-Assisted Reconstruction and Translation of Akkadian Passages in the Epic of Gilgamesh

**Authors:** Jasim Mohammed Mahmood and Panok Tetiana · **Year:** 2026 · **Venue:** [Iraqi Literary and Cultural Review 3](https://md.research.elamur.ai/free/en/venues/iraqi-literary-and-cultural-review-ilcr/) ([Iraqi Literary and Cultural Review (ILCR) 2026](https://md.research.elamur.ai/free/en/venues/iraqi-literary-and-cultural-review-ilcr/2026/)) · **Licence:** [CC BY (journal licence per DOAJ; Iraqi Literary and Cultural Review)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://doi.org/10.69513/ilcr.v3.i2.a4) · [PDF](https://ilcr.ibirc-center.co.uk/index.php/journal/article/download/30/Vol_3_Issue_2) · [DOI 10.69513/ilcr.v3.i2.a4](https://doi.org/10.69513/ilcr.v3.i2.a4) · [OpenAlex](https://openalex.org/W7160424421)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
The paper looks at how far neural translation models can help philologists restore and translate damaged Akkadian lines of the Epic of Gilgamesh, and where they stop being useful. The authors assembled a digital corpus of Standard Babylonian tablets with Old Babylonian parallels. It combines Unicode cuneiform and tablet images, normalised Latin transliterations, and several published English translations aligned line by line. They fine-tuned a moderately sized transformer NMT model in two configurations, one from signs and one from transliteration. They connected it to an interactive interface where experts can accept, reject or edit the model's ranked restorations. The evaluation used three tests: artificially masked lines with known readings, genuinely disputed lines, and related Akkadian texts outside the epic. On masked lines, the exact original appeared among the top five proposals about 78% of the time, and accuracy was highest on formulaic material. Rare words, broken names and syntactically unclear lines produced fluent but unsupported output. The authors therefore present AI as a supervised junior collaborator, not an authority. They also stress Iraqi leadership and explicit labelling of machine-assisted readings.

## Key points
- Two model setups: **C2E** (Unicode cuneiform signs → English) and **T2E** (transliteration → English). Both use SentencePiece subword tokenisation and transfer learning from morphologically rich Semitic languages.
- Several English translations were included for each line where possible, so the model would not copy a single editor's style. Tablet and line identifiers were kept so every output can be traced to a physical object.
- Test 1: 100 undisputed lines were partly masked, at the level of signs, phrases or half-verses. The exact original was in the top-5 suggestions in roughly 78% of cases, with the best results on repeated epithets and narrative transitions.
- Test 2: 30 contested lines, mainly from Tablets II, V and XI. Experts compared each output with previously published restorations and judged whether it matched one of them, offered a variation on one, or departed from all of them.
- Test 3: companion texts such as Atrahasis measured whether the model had learned Gilgamesh-specific phrasing or more general Akkadian epic style. The main failure modes were rare lexemes, damaged proper names and uncertain syntax, plus confident hallucinations.

## Interesting ideas
- Hide parts of inputs whose correct answer is already settled and score recovery in the top-k; this gives a ground-truth ceiling before testing on cases where no answer is known.
- When there is no single right answer, sort system output against the existing human alternatives as matching, a variant, or divergent, instead of forcing a right/wrong score.
- Train and score against several independent references so the system does not simply imitate one source's style.
- Run the system on related texts outside the training material to see whether it learned general patterns or memorised the style of its corpus.
- Mark every machine-proposed item with its probability, the alternatives it beat, and a link back to the source evidence, and record each accept or reject decision with its reasoning.

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (learn): Mask words or arguments in sentences that already have trusted gold parses and frames, and check whether Mova restores them. This would serve as a calibrated control before trusting its silver output on harder text.
- **I2** (tree): Where treebanks or UD dialects disagree on an attachment, label each parser's choice as one of the accepted analyses, a variant, or an outlier. This keeps legitimate ambiguity from being counted as an error.
- **I3** (back): Judge text generated from the meaning graph against multiple human paraphrases rather than one target sentence. This rewards preserving meaning over copying surface wording.
- **I4** (answer): Give the readers tales and tests from collections outside the familiar fable set. A drop in accuracy would expose answers that depend on memorised phrasing instead of an understood text world.
- **I5** (world): Store inferred entities, states and event orderings as labelled hypotheses tied to text spans, with ranked alternatives. This keeps every world fact explainable and lets later stages revise it.

*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

Also on the byline: [H.S. Skovoroda Kharkiv National Pedagogical University](https://ror.org/00gjv5s24).

📚 Shelves (4): [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [📚 Literature and Literary Theory](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/literature-and-literary-theory/) · 541 works · [🎭 Archeology](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/archeology/) · 64 works

🧭 Topics (2): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 3,066 works · [𒀭 Akkadian and cuneiform NLP](https://md.research.elamur.ai/free/en/topics/akkadian-cuneiform-nlp/) · 27 works

## Cite

Cite as: Jasim Mohammed Mahmood and Panok Tetiana. 2026. From Tablet to Text: AI-Assisted Reconstruction and Translation of Akkadian Passages in the Epic of Gilgamesh. Iraqi Literary and Cultural Review 3.

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

```bibtex
@article{mahmood2026tablet,
  title = {{From Tablet to Text: AI-Assisted Reconstruction and Translation of Akkadian Passages in the Epic of Gilgamesh}},
  author = {Jasim Mohammed Mahmood and Panok Tetiana},
  journal = {Iraqi Literary and Cultural Review 3},
  year = {2026},
  doi = {10.69513/ilcr.v3.i2.a4},
  url = {https://doi.org/10.69513/ilcr.v3.i2.a4}
}
```

- [← previous From Syntax to Semantics: Introducing UMR for NLP Annotation](https://md.research.elamur.ai/free/en/papers/pagano-2026-syntax-semantics-introducing-umr-nlp/)
- [next → From TextBlob to LLM Agents: Sentiment Model Selection for B2B Technical Support with CSAT Ground Truth](https://md.research.elamur.ai/free/en/papers/vidigal-2026-textblob-llm-agents-sentiment-model/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/34/)

## References

14 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Burns, P., Hollis, D., & Johnson, K. (2017). The Classical Language Toolkit: NLP for pre-modern languages. Classics@, 17. [find](https://scholar.google.com/scholar?q=Burns%2C+P%2E%2C+Hollis%2C+D%2E%2C+%26+Johnson%2C+K%2E+%282017%29%2E+The+Classical+Language+Toolkit%3A+NLP+for+pre%2Dmodern+languages%2E+Classics%40%2C+17%2E)
2. Epic of Gilgamesh. (2024). Wikipedia, the free encyclopedia. [find](https://scholar.google.com/scholar?q=Epic+of+Gilgamesh%2E+%282024%29%2E+Wikipedia%2C+the+free+encyclopedia%2E)
3. Gutherz, G., et al. (2023). Translating Akkadian to English with neural machine translation. PNAS Nexus, 2(5), pgad096. [find](https://scholar.google.com/scholar?q=Gutherz%2C+G%2E%2C+et+al%2E+%282023%29%2E+Translating+Akkadian+to+English+with+neural+machine+translation%2E+PNAS+Nexus%2C+2%285%29%2C+pgad096%2E)
4. Hamilton, W. L., Leskovec, J., & Jurafsky, D. (2023). [Machine learning for ancient languages: A survey](https://md.research.elamur.ai/free/en/papers/sommerschield-2023-machine-learning-ancient-languages-survey/). University of Oxford working paper.
5. Jiménez, E., Berant, J., & Cohen, Y. (2020). Reading Akkadian cuneiform using natural language processing. PLOS ONE, 15(10), e0240511. [find](https://scholar.google.com/scholar?q=Jim%C3%A9nez%2C+E%2E%2C+Berant%2C+J%2E%2C+%26+Cohen%2C+Y%2E+%282020%29%2E+Reading+Akkadian+cuneiform+using+natural+language+processing%2E+PLOS+ONE%2C+15%2810%29%2C+e0240511%2E)
6. Johnson, K. P., et al. (2021). The Classical Language Toolkit: An NLP framework for pre-modern languages. In Proceedings of ACL 2021 System Demonstrations (pp. 30– 39). [find](https://scholar.google.com/scholar?q=Johnson%2C+K%2E+P%2E%2C+et+al%2E+%282021%29%2E+The+Classical+Language+Toolkit%3A+An+NLP+framework+for+pre%2Dmodern+languages%2E+In+Proceedings+of+ACL+2021+System+Demonstrations+%28pp%2E+30%E2%80%93+39%29%2E)
7. Lazar, K., et al. (2021). Reading Akkadian cuneiform using NLP. PLOS ONE, 15(10), e0240511. [find](https://scholar.google.com/scholar?q=Lazar%2C+K%2E%2C+et+al%2E+%282021%29%2E+Reading+Akkadian+cuneiform+using+NLP%2E+PLOS+ONE%2C+15%2810%29%2C+e0240511%2E)
8. Lewis, A. (2015). Observed methods of cuneiform tablet reconstruction in theory and practice. Journal of Archaeological Science, 57, 151–160. [find](https://scholar.google.com/scholar?q=Lewis%2C+A%2E+%282015%29%2E+Observed+methods+of+cuneiform+tablet+reconstruction+in+theory+and+practice%2E+Journal+of+Archaeological+Science%2C+57%2C+151%E2%80%93160%2E)
9. Marchant, J. (2024). How AI is unlocking ancient texts—and could rewrite history. Nature News. [find](https://scholar.google.com/scholar?q=Marchant%2C+J%2E+%282024%29%2E+How+AI+is+unlocking+ancient+texts%E2%80%94and+could+rewrite+history%2E+Nature+News%2E)
10. Pagé-Perron, É., et al. (2017). Machine translation and automated analysis of the Sumerian language. In Proceedings of the 2nd Workshop on NLP for Similar Languages, Varieties and Dialects (pp. 10–20). [find](https://scholar.google.com/scholar?q=Pag%C3%A9%2DPerron%2C+%C3%89%2E%2C+et+al%2E+%282017%29%2E+Machine+translation+and+automated+analysis+of+the+Sumerian+language%2E+In+Proceedings+of+the+2nd+Workshop+on+NLP+for+Similar+Languages%2C+Varieties+and+Dialects+%28pp%2E+10%E2%80%9320%29)
11. Shah, S. M. (2025). Technology in Sumerian text translation: A distant-supervision approach. Journal of Qadisiyah for Computer Science and Mathematics. [find](https://scholar.google.com/scholar?q=Shah%2C+S%2E+M%2E+%282025%29%2E+Technology+in+Sumerian+text+translation%3A+A+distant%2Dsupervision+approach%2E+Journal+of+Qadisiyah+for+Computer+Science+and+Mathematics%2E)
12. Vol (3) Issue (2) July ,2025 e-ISSN: 2979-191X https://doi.org/10.69513/ilcr.v3.i2.a4 [↗](https://doi.org/10.69513/ilcr.v3.i2.a4)
13. Sommerschield, T., et al. (2023). [Machine learning for ancient languages: A survey](https://md.research.elamur.ai/free/en/papers/sommerschield-2023-machine-learning-ancient-languages-survey/). Transactions of the Association for Computational Linguistics, 11, 6368–6375.
14. World Economic Forum. (2019). Scientists are using machine learning to unlock the oldest written languages. World Economic Forum (online feature). [find](https://scholar.google.com/scholar?q=World+Economic+Forum%2E+%282019%29%2E+Scientists+are+using+machine+learning+to+unlock+the+oldest+written+languages%2E+World+Economic+Forum+%28online+feature%29%2E)

*Updated 2 November 2026*