---
title: "A Lightweight N-gram Approach to Abbreviation Expansion in"
description: "The authors propose an unsupervised method that picks expansions using only N-gram statistics from a large in-domain corpus… 9 references."
authors: ["Tjaša Šoltes", "Marko Bajec"]
year: 2026
venue: "Proceedings of the Third Workshop on Computation and Written Language (CAWL 2026) @ LREC 2026"
source: "https://aclanthology.org/2026.cawl-1.10/"
doi: "10.63317/3t4xjiz2vvw2"
licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)"
licence_url: "https://creativecommons.org/licenses/by-nc/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/soltes-2026-lightweight-n-gram-approach-abbreviation/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/soltes-2026-lightweight-n-gram-approach-abbreviation/)

[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/20/) / A Lightweight N-gram Approach to Abbreviation Expansion in Large Corpora

# A Lightweight N-gram Approach to Abbreviation Expansion in Large Corpora

**Authors:** Tjaša Šoltes and Marko Bajec · **Year:** 2026 · **Venue:** [Proceedings of the Third Workshop on Computation and Written Language (CAWL 2026) @ LREC 2026](https://md.research.elamur.ai/free/en/venues/cawl/) ([Workshop on Computation and Written Language @ LREC 2026](https://md.research.elamur.ai/free/en/venues/cawl/2026/)) · **Publisher:** [European Language Resources Association (ELRA)](https://md.research.elamur.ai/free/en/publishers/european-language-resources-association-elra/) · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.cawl-1.10/) · [PDF](https://aclanthology.org/2026.cawl-1.10.pdf) · [DOI 10.63317/3t4xjiz2vvw2](https://doi.org/10.63317/3t4xjiz2vvw2) · [OpenAlex](https://openalex.org/W7166563604)

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

## Summary
In pipelines such as TTS and ASR, abbreviations must be converted to full words. In Slovene this is harder because the full form must also carry the correct grammatical case and number. The authors propose an unsupervised method that picks expansions using only N-gram statistics from a large in-domain corpus, with no dictionaries, morphological analysers or pretrained models. They use a proprietary Slovene medical corpus of about 59 million tokens and treat any letter-only token ending in a dot as an abbreviation. Candidate expansions come from five bigram/trigram context tables, stored as compact integer-indexed binary files, and are filtered by a shared prefix. Candidates are then ranked in one of two ways:
- **Strict backoff:** use the single best available context.
- **Linear interpolation:** combine all contexts, scaling each one by a reliability factor that dampens sparse evidence.

Reliability-aware interpolation was the most robust overall, especially for medium- and low-frequency abbreviations, while backoff stayed competitive for very rare ones. Excluding left context from candidate generation also raised accuracy. The work shows that surface statistics can implicitly handle inflection, and the method runs quickly on ordinary hardware.

## Key points
- **Corpus:** proprietary Slovene medical corpus with 3,893,434 sentences and 59,083,266 tokens (137,423 types). It contains 237 abbreviation types with 87,610 occurrences in total, and each abbreviation has a median of 4 possible expansions. The authors list 34 conventional abbreviations (e.g., *dr.*, *npr.*, *b. p.*) to exclude from expansion, and found 18 false abbreviations caused by sentence-splitting errors.
- **Context types:** center trigram (C3), left and right trigrams (L3/R3), and left and right bigrams (L2/R2).
- **Interpolation score:** each context's normalized candidate count is weighted by λ_i and multiplied by the reliability factor R_i = N_i/(N_i + k).
- **Evaluation:** accuracy is reported across five frequency bands, from very frequent (>5,000 occurrences) to very rare (\<10). The weakest results are on very rare abbreviations and on orthographically similar candidates.
- **Efficiency:** the binary tables take 169.8 MB. On a CPU-only laptop (AMD Ryzen 7 5700U, 16 GB RAM), the system processes about 2,375 abbreviation instances per second.

*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

- [University of Ljubljana](https://md.research.elamur.ai/free/en/universities/university-of-ljubljana/) — University, Ljubljana, Slovenia · 13 works · 38,693 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 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

## Cite

Cite as: Tjaša Šoltes and Marko Bajec. 2026. A Lightweight N-gram Approach to Abbreviation Expansion in Large Corpora. Proceedings of the Third Workshop on Computation and Written Language (CAWL 2026) @ LREC 2026.

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

```bibtex
@inproceedings{soltes2026lightweight,
  title = {{A Lightweight N-gram Approach to Abbreviation Expansion in Large Corpora}},
  author = {Tjaša Šoltes and Marko Bajec},
  booktitle = {Proceedings of the Third Workshop on Computation and Written Language (CAWL 2026) @ LREC 2026},
  year = {2026},
  doi = {10.63317/3t4xjiz2vvw2},
  url = {https://doi.org/10.63317/3t4xjiz2vvw2}
}
```

- [← previous A Lexicon-Grammar of Brazilian Portuguese Predicative Adjectives](https://md.research.elamur.ai/free/en/papers/martinez-2026-lexicon-grammar-brazilian-portuguese-predicative/)
- [next → A Linguistic Ontology for Constructicography: The Research Constructicon and its Ontology Modules](https://md.research.elamur.ai/free/en/papers/winckel-2026-linguistic-ontology-constructicography-research-constructicon/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/20/)

## References

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

1. Daphné Chopard and Irena Spasić. 2019. A deep learning approach to self-expansion of abbreviations based on morphology and context distance. In Statistical Language and Speech Processing: 7th International Conference, SLSP 2019, Ljubljana, Slovenia, October 14–16, 2019, Proceedings, page 71–82, Berl… [find](https://scholar.google.com/scholar?q=Daphn%C3%A9+Chopard+and+Irena+Spasi%C4%87%2E+2019%2E+A+deep+learning+approach+to+self%2Dexpansion+of+abbreviations+based+on+morphology+and+context+distance%2E+In+Statistical+Language+and+Speech+Processing%3A+7th+Internat)
2. Paul Cook and Suzanne Stevenson. 2009. An unsupervised model for text message normalization. In Proceedings of the Workshop on Computational Approaches to Linguistic Creativity, pages 71–78, Boulder, Colorado. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Paul+Cook+and+Suzanne+Stevenson%2E+2009%2E+An+unsupervised+model+for+text+message+normalization%2E+In+Proceedings+of+the+Workshop+on+Computational+Approaches+to+Linguistic+Creativity%2C+pages+71%E2%80%9378%2C+Boulder%2C+)
3. Kyle Gorman, Christo Kirov, Brian Roark, and Richard Sproat. 2021. Structured abbreviation expansion in context. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 995–1005, Punta Cana, Dominican Republic. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Kyle+Gorman%2C+Christo+Kirov%2C+Brian+Roark%2C+and+Richard+Sproat%2E+2021%2E+Structured+abbreviation+expansion+in+context%2E+In+Findings+of+the+Association+for+Computational+Linguistics%3A+EMNLP+2021%2C+pages+995%E2%80%93100)
4. Amila Kugic, Stefan Schulz, and Markus Kreuzthaler. 2024. Disambiguation of acronyms in clinical narratives with large language models. Journal of the American Medical Informatics Association, 31(9):2040–2046. [find](https://scholar.google.com/scholar?q=Amila+Kugic%2C+Stefan+Schulz%2C+and+Markus+Kreuzthaler%2E+2024%2E+Disambiguation+of+acronyms+in+clinical+narratives+with+large+language+models%2E+Journal+of+the+American+Medical+Informatics+Association%2C+31%289%29%3A2)
5. Nima Shafiei Rezvani Nezhad, Meysam Mansouri, Rabih Abdulkarim Zakaria, and Ruhollah Abolhasani. 2025. Medical abbreviation disambiguation with large language models: Zero- and few-shot evaluation on the medal dataset. bioRxiv. [find](https://scholar.google.com/scholar?q=Nima+Shafiei+Rezvani+Nezhad%2C+Meysam+Mansouri%2C+Rabih+Abdulkarim+Zakaria%2C+and+Ruhollah+Abolhasani%2E+2025%2E+Medical+abbreviation+disambiguation+with+large+language+models%3A+Zero%2D+and+few%2Dshot+evaluation+on+)
6. Serguei V. S. Pakhomov, Ted Pedersen, and Christopher G. Chute. 2005. Abbreviation and acronym disambiguation in clinical discourse. AMIA ... Annual Symposium proceedings. AMIA Symposium, pages 589–93. [find](https://scholar.google.com/scholar?q=Serguei+V%2E+S%2E+Pakhomov%2C+Ted+Pedersen%2C+and+Christopher+G%2E+Chute%2E+2005%2E+Abbreviation+and+acronym+disambiguation+in+clinical+discourse%2E+AMIA+%2E%2E%2E+Annual+Symposium+proceedings%2E+AMIA+Symposium%2C+pages+589%E2%80%9393)
7. Alvin Rajkomar, Eric Loreaux, Yuchen Liu, Jonas Kemp, Benny Li, Ming-Jun Chen, Yi Zhang, Afroz Mohiuddin, Juraj Gottweis, et al. 2022. Deciphering clinical abbreviations with a privacy protecting machine learning system. Nature Communications, 13(1):7456. [find](https://scholar.google.com/scholar?q=Alvin+Rajkomar%2C+Eric+Loreaux%2C+Yuchen+Liu%2C+Jonas+Kemp%2C+Benny+Li%2C+Ming%2DJun+Chen%2C+Yi+Zhang%2C+Afroz+Mohiuddin%2C+Juraj+Gottweis%2C+et+al%2E+2022%2E+Deciphering+clinical+abbreviations+with+a+privacy+protecting+mach)
8. Brian Roark and Richard Sproat. 2014. Hippocratic abbreviation expansion. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pages 364–369, Baltimore, Maryland. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Brian+Roark+and+Richard+Sproat%2E+2014%2E+Hippocratic+abbreviation+expansion%2E+In+Proceedings+of+the+52nd+Annual+Meeting+of+the+Association+for+Computational+Linguistics+%28Volume+2%3A+Short+Papers%29%2C+pages+364)
9. Akira Terada, Takenobu Tokunaga, and Hozumi Tanaka. 2004. Automatic expansion of abbreviations by using context and character information. Information Processing Management, 40(1):31– 45. [find](https://scholar.google.com/scholar?q=Akira+Terada%2C+Takenobu+Tokunaga%2C+and+Hozumi+Tanaka%2E+2004%2E+Automatic+expansion+of+abbreviations+by+using+context+and+character+information%2E+Information+Processing+Management%2C+40%281%29%3A31%E2%80%93+45%2E)

*Updated 1 November 2026*