The Mediomatix Corpus: Parallel Data for Romansh Language Varieties via Comparable Schoolbooks

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hopton, Zachary, Vamvas, Jannis, Büchler, Andrin, Rutkiewicz, Anna, Cathomas, Rico, Sennrich, Rico
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908830709841920
author Hopton, Zachary
Vamvas, Jannis
Büchler, Andrin
Rutkiewicz, Anna
Cathomas, Rico
Sennrich, Rico
author_facet Hopton, Zachary
Vamvas, Jannis
Büchler, Andrin
Rutkiewicz, Anna
Cathomas, Rico
Sennrich, Rico
contents The five idioms (i.e., varieties) of the Romansh language are largely standardized and are taught in the schools of the respective communities in Switzerland. In this paper, we present the first parallel corpus of Romansh idioms. The corpus is based on 291 schoolbook volumes, which are comparable in content for the five idioms. We use automatic alignment methods to extract 207k multi-parallel segments from the books, with more than 2M tokens in total. A small-scale human evaluation confirms that the segments are highly parallel, making the dataset suitable for NLP applications such as machine translation between Romansh idioms. We release the parallel and unaligned versions of the dataset under a CC-BY-NC-SA license and demonstrate its utility for machine translation by training and evaluating an LLM and a supervised multilingual MT model on the dataset.
format Preprint
id arxiv_https___arxiv_org_abs_2508_16371
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle The Mediomatix Corpus: Parallel Data for Romansh Language Varieties via Comparable Schoolbooks
Hopton, Zachary
Vamvas, Jannis
Büchler, Andrin
Rutkiewicz, Anna
Cathomas, Rico
Sennrich, Rico
Computation and Language
The five idioms (i.e., varieties) of the Romansh language are largely standardized and are taught in the schools of the respective communities in Switzerland. In this paper, we present the first parallel corpus of Romansh idioms. The corpus is based on 291 schoolbook volumes, which are comparable in content for the five idioms. We use automatic alignment methods to extract 207k multi-parallel segments from the books, with more than 2M tokens in total. A small-scale human evaluation confirms that the segments are highly parallel, making the dataset suitable for NLP applications such as machine translation between Romansh idioms. We release the parallel and unaligned versions of the dataset under a CC-BY-NC-SA license and demonstrate its utility for machine translation by training and evaluating an LLM and a supervised multilingual MT model on the dataset.
title The Mediomatix Corpus: Parallel Data for Romansh Language Varieties via Comparable Schoolbooks
topic Computation and Language
url https://arxiv.org/abs/2508.16371