A multi-lingual corpus in German, Romansh and Italian
The parallel corpus is presented as JSON files, one per year.
Tab seperated files containing parallel sentences in the three combinations: German-Romansh (DE-RM), German-Italian (DE-IT) and Romansh-Italian (RM-IT).
| Combination | Sentences | Tokens | Types |
|---|---|---|---|
| DE-RM | 106091 | 1864886, 2377547 | 97385, 50869 |
| DE-IT | 103441 | 1848214, 2219378 | 96859, 58050 |
| RM-IT | 102757 | 2302374, 2165287 | 50014, 57642 |
The copyright to the press releases lies with the Canton of Graubünden. Downloading the press releases is explicitly allowed. Commercial use is explicity prohibited. In any event, please consult the copyright notice on the cantnon's website.
A set of 600 parallel sentences (German-Romansh), manually annotated for word alignment.
If you use this corpus or the gold standard for published work, please cite my paper:
@inproceedings{dolev-2023-mbert,
title = "Does m{BERT} understand {R}omansh? Evaluating word embeddings using word alignment.",
author = "Dolev, Eyal Liron",
editor = {Ghorbel, Hatem and
Sokhn, Maria and
Cieliebak, Mark and
H{\"u}rlimann, Manuela and
de Salis, Emmanuel and
Guerne, Jonathan},
booktitle = "Proceedings of the 8th edition of the Swiss Text Analytics Conference",
month = jun,
year = "2023",
address = "Neuchatel, Switzerland",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.swisstext-1.5",
pages = "41--53",
}