# Wdowiak customizes NLLB for sicilian translation

Digest AI · Generative AI & Models · published 2026-10-23T04:00:00Z

Canonical: https://digestai.news/story/wdowiak-customizes-nllb-for-sicilian-translation

## Summary

Wdowiak rebuilt a small neural machine translation system for Sicilian by extracting parallel text from the bilingual journal Arba Sicula and porting the original tokenization to a Perl‑free implementation. The new system was evaluated against a frozen test set from the journal.

The study compared small Transformers trained from scratch with the pretrained NLLB model adapted by low‑rank fine‑tuning. The pretrained model performed better overall, and staged training—first forward‑scored back‑translations of web Sicilian, then curated literary text—improved translation into Sicilian more than reverse training.

On the test set, the staged model with a larger adapter and more back‑translations outperformed the reverse‑training recipe in English‑related directions but lagged between Sicilian and Italian. A multilingual adapter degraded performance on unseen language pairs, limiting its general use.

## Key points

- Wdowiak rebuilt Sicilian NMT using NLLB with low‑rank fine‑tuning
- Staged training with back‑translations improved Sicilian translation more than reverse training
- Multilingual adapter degraded performance on unseen language pairs

## Why it matters

The work shows that adapting large multilingual models with low‑rank fine‑tuning and staged training can yield better performance on low‑resource languages like Sicilian, offering a path for preserving linguistic diversity and improving translation services for minority languages.

## Sources

1. [Climbing the Ladder: Automating Data Pipelines and Reproducing Baselines in Low-Resource Settings](https://arxiv.org/abs/2610.30374) (arXiv cs.CL, 2026-10-23, primary source)

## Cite

Digest AI, "Wdowiak customizes NLLB for sicilian translation", 23 October 2026, https://digestai.news/story/wdowiak-customizes-nllb-for-sicilian-translation

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/wdowiak-customizes-nllb-for-sicilian-translation.json