Parallel Monolingual Corpora: Difference between revisions
No edit summary |
No edit summary |
||
| Line 36: | Line 36: | ||
<!--T:12--> | <!--T:12--> | ||
1) The first dataset is created by Bram Vanroy for Dutch text simplification tasks using text-to-text transfer transformers.It comprises Dutch source sentences along with their corresponding simplified sentences, generated with ChatGPT. | 1) The first dataset is created by Bram Vanroy for Dutch text simplification tasks using text-to-text transfer transformers. It comprises Dutch source sentences along with their corresponding simplified sentences, generated with ChatGPT. | ||
<!--T:13--> | <!--T:13--> | ||
| Line 47: | Line 47: | ||
<!--T:15--> | <!--T:15--> | ||
2) The second dataset is created by UWV Nederland as part of the "Leesplank" project to | 2) The second dataset is created by UWV Nederland as part of the "Leesplank" project, which is an effort to generate datasets that are ethically and legally sound. The dataset comprises 2.87 million paragraphs and its simplified text as corresponding result. The paragraphs are based on the Dutch Wikipedia extract from [http://gigacorpus.nl/ Gigacorpus]. The text was filtered and cleaned by [https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/content-filter?tabs=warning%2Cpython-new using GPT-4 1106 preview]. | ||
<!--T:16--> | <!--T:16--> | ||
| Line 54: | Line 54: | ||
<!--T:17--> | <!--T:17--> | ||
A more extended version of this dataset was made by Michiel Buisman and Bram Vanroy. This datasets contains a first, small set of variations of Wikipedia paragraphs in different styles (jargon, official, | A more extended version of this dataset was made by Michiel Buisman and Bram Vanroy. This datasets contains a first, small set of variations of Wikipedia paragraphs in different styles (jargon, official, archaïc language, technical, academic, and poetic). | ||
<!--T:18--> | <!--T:18--> | ||
| Line 61: | Line 61: | ||
<!--T:19--> | <!--T:19--> | ||
3) The third dataset is the comparable corpus created by Nick Vanackere. It contains | 3) The third dataset is the comparable corpus created by Nick Vanackere. It contains 12,687 Wablieft articles between 2012-2017 and 206,466 De Standaard articles from 2013-2017. To ensure comparability, only articles from 08/01/2013 till 16/11/2017 were considered, resulting in 8,744 Wablieft articles and 202,284 De Standaard articles. The difference in the number of articles is due to the publication frequency: Wablieft is published once a week and the Standaard daily. | ||
<!--T:20--> | <!--T:20--> | ||
Revision as of 15:15, 11 June 2024
DAESO Corpus
The DAESO Corpus is a parallel monolingual treebank of Dutch texts and the corpus contains more than 2.1 million words of parallel and comparable text. About 678,000 words were lined up manually and about 1.5 million words were automatically aligned. A semantic relation was added to the aligned words / phrases.
- 92.5 MB
- version 1.0 (2010)
- Download page
Bible Corpus
A diachronically and synchronically parallel corpus of Bible translations in Dutch, English, German and Swedish, with texts from the 14th century until today.
Simplification Data
Manually created datasets
The Dutch municipal corpus is a parallel monolingual corpus for the evaluation of sentence-level simplification in the Dutch municipal domain. The corpus was created by Amsterdam Intelligence. It contains 1,311 translated parallel sentence pairs that were automatically aligned. The sentence pairs originate from 50 documents from the Communications Department of the City of Amsterdam that were manually simplified to evaluate simplification for Dutch.
- 265 KB
- Download dataset (CSV file)
Automatically created datasets
1) The first dataset is created by Bram Vanroy for Dutch text simplification tasks using text-to-text transfer transformers. It comprises Dutch source sentences along with their corresponding simplified sentences, generated with ChatGPT.
- Training = 1013 sentences (262 KB)
- Validation = 126 sentences (32.6 KB)
- Test = 128 sentences (33 KB)
2) The second dataset is created by UWV Nederland as part of the "Leesplank" project, which is an effort to generate datasets that are ethically and legally sound. The dataset comprises 2.87 million paragraphs and its simplified text as corresponding result. The paragraphs are based on the Dutch Wikipedia extract from Gigacorpus. The text was filtered and cleaned by using GPT-4 1106 preview.
- 3.02 MB
- Download page
A more extended version of this dataset was made by Michiel Buisman and Bram Vanroy. This datasets contains a first, small set of variations of Wikipedia paragraphs in different styles (jargon, official, archaïc language, technical, academic, and poetic).
- 3.02 MB
- Download page
3) The third dataset is the comparable corpus created by Nick Vanackere. It contains 12,687 Wablieft articles between 2012-2017 and 206,466 De Standaard articles from 2013-2017. To ensure comparability, only articles from 08/01/2013 till 16/11/2017 were considered, resulting in 8,744 Wablieft articles and 202,284 De Standaard articles. The difference in the number of articles is due to the publication frequency: Wablieft is published once a week and the Standaard daily.
- 17.5 MB
Translated datasets
1 The first translated dataset is created by Netherlands Forensic Institute using Meta's No Language Left Behind model. It comprises 167,000 aligned sentence pairs and serves as a Dutch translation of the SimpleWiki dataset.
- 8.67 MB
- Download dataset
2 The second translated dataset is created by Theresa Seidl in the context of Controllable sentence simplification in Dutch. This is a synthetic dataset which is a combination of the first 10,000 rows of the parallel WikiLarge dataset, and ASSET (Abstractive Sentence Simplification Evaluation and Tuning) dataset. By combining these two datasets, Theresa translated them to Dutch using Google Neural Machine Translation.