Akkadian and cuneiform NLP

15 paper cards from Mova's library

What it is

Akkadian is the East Semitic language of ancient Mesopotamia, spoken and written for more than two thousand years in its Babylonian and Assyrian dialects. It was a language of empire and of letters: treaties, laws, omens, epics and everyday contracts survive on hundreds of thousands of clay tablets.

Cuneiform is the script those tablets use: wedge-shaped signs pressed into wet clay with a reed. It was first made for Sumerian and later carried Akkadian, Hittite, Elamite and others. One sign can stand for a whole word or for a syllable, and most signs have several readings, so reading a tablet means choosing a reading for every sign.

Cuneiform NLP turns that into tasks a machine can help with: transliterating signs, normalizing the transliteration into words, lemmatizing and analyzing the root-and-pattern verb, restoring broken passages and translating. The data is small and damaged, which is why rules and careful counting still matter here.

How Mova does it

Mova does not read Akkadian yet. It keeps the work on it on its Semitic shelf because Akkadian bears directly on the languages it does read: Biblical Aramaic took many words from Akkadian through the Assyrian and Babylonian administration, the oldest Aramaic inscriptions include an Akkadian–Aramaic bilingual, and the three languages share the root system that Mova's Hebrew and Aramaic segmenters are built around.

• The Aramaic lexicon, with its Akkadian loanwords
• The Aramaic family and its inscriptions
• The Semitic shelf of the paper cards
• How the paper cards are made

The papers (15)

Cards with a measured or kept idea first, then by citations.

1. Neural Decipherment via Minimum-Cost Flow: From Ugaritic to Linear B — Jiaming Luo, Yuan Cao, Regina Barzilay, 2019 · 33 citations
2. Filling the Gaps in Ancient Akkadian Texts: A Masked Language Modelling Approach — Koren Lazar, Benny Saret, Asaf Yehudai, Wayne Horowitz et al., 2021 · 18 citations
3. Akkadian Treebank for early Neo-Assyrian Royal Inscriptions — Mikko Luukko, Aleksi Sahala, Sam Hardwick, Krister Lindén, 2020 · 6 citations
4. Parsing Akkadian Verbs with Prolog — Aaron Macks, 2002 · 6 citations
5. A Morphological Analyzer for Akkadian Verbal Forms with a Model of Phonetic Transformations — Francois Barthelemy, 1998 · 5 citations
6. Opuscula Ugaritico-Accadico-Hebraica: Relative particles, paʿam, and Amraphel — Ola Wikander, 2019 · 1 citations
7. Une description morphologique structurée en arbre du verbe akkadien qui utilise des structures de traits et des transducteurs multirubans [A Tree-Structured morphological description of the Akkadian verb which uses Feature Structures and Multi-tape Transducers] — François Barthélémy, 2009 · 1 citations
8. TabletCraft: Bridging a 4,000-Year Cultural Gap with Bidirectional Akkadian NMT and Cuneiform Rendering — Zhaohui Geoffrey Wang, 2026
9. Beyond Base Predictors: Using LLMs to Resolve Ambiguities in Akkadian Lemmatization — Frederick Riemenschneider, 2025
10. EvaCun 2025 Shared Task: Lemmatization and Token Prediction in Akkadian and Sumerian using LLMs — Shai Gordin, Aleksi Sahala, Shahar Spencer, Stav Klein, 2025
11. Loanwords in the Language of Archaic Biblical Hebrew Poetry: A Case for Akkadianisms — Tania Notarius, 2025
12. A First Look at the Ugaritic Poetic Text Corpus — Tillmann Dönicke, Clemens Steinberger, Max-Ferdinand Zeterberg, Noah Kröll, 2024
13. At the Crossroad of Cuneiform and NLP: Challenges for Fine-grained Part-of-speech Tagging — Gustav Ryberg Smidt, Els Lefever, Katrien de Graef, 2024
14. Automated Phonological Transcription of Akkadian Cuneiform Text — Aleksi Sahala, Miikka Silfverberg, Antti Arppe, Krister Lindén, 2020
15. Word Segmentation for Akkadian Cuneiform — Timo Homburg, Christian Chiarcos, 2016

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/topics/akkadian-cuneiform-nlp/