|
--- |
|
license: cc-by-4.0 |
|
language: |
|
- fr |
|
- en |
|
- it |
|
- de |
|
task_categories: |
|
- text-classification |
|
task_ids: |
|
- natural-language-inference |
|
- multi-input-text-classification |
|
size_categories: |
|
- 1K<n<10K |
|
--- |
|
# Dataset Card for Dataset Name |
|
|
|
## Dataset Description |
|
|
|
- **Homepage:** |
|
- **Repository:** https://github.com/mskandalis/rte3-french |
|
- **Paper:** |
|
- **Leaderboard:** |
|
- **Point of Contact:** |
|
|
|
### Dataset Summary |
|
|
|
This repository contains all manually translated versions of RTE-3 dataset, plus the original English one. The languages into which RTE-3 dataset has so far been translated are Italian (2012), German (2013), and French (2023). |
|
|
|
Unlike in other repositories, both our own French version and the older Italian and German ones are here annotated in 3 classes (entailment, neutral, contradiction), and not in 2 (entailment, not entailment). |
|
|
|
If you want to use the dataset only in a specific language among those provided here, you can filter data by selecting only the language column value you wish. |
|
|
|
### Supported Tasks and Leaderboards |
|
|
|
This dataset can be used for the task of Natural Language Inference (NLI), also known as Recognizing Textual Entailment (RTE), which is a sentence-pair classification task. |
|
|
|
## Dataset Structure |
|
|
|
### Data Fields |
|
|
|
- `id`: Index number. |
|
- `language`: The language of the concerned pair of sentences. |
|
- `premise`: The translated premise in the target language. |
|
- `hypothesis`: The translated premise in the target language. |
|
- `label`: The classification label, with possible values 0 (`entailment`), 1 (`neutral`), 2 (`contradiction`). |
|
- `label_text`: The classification label, with possible values `entailment` (0), `neutral` (1), `contradiction` (2). |
|
- `task`: The particular NLP task that the data was drawn from (IE, IR, QA and SUM). |
|
- `length`: The length of the text of the pair. |
|
|
|
### Data Splits |
|
|
|
| name |development|test| |
|
|-------------|----------:|---:| |
|
|all_languages| 3200 |3200| |
|
| fr | 800 | 800| |
|
| de | 800 | 800| |
|
| it | 800 | 800| |
|
| en | 800 | 800| |
|
|
|
For French RTE-3: |
|
| name |entailment|neutral|contradiction| |
|
|-------------|---------:|------:|------------:| |
|
| dev | 412 | 299 | 89 | |
|
| test | 410 | 318 | 72 | |
|
|
|
| name |short|long| |
|
|-------------|----:|---:| |
|
| dev | 665 | 135| |
|
| test | 683 | 117| |
|
|
|
| name | IE| IR| QA|SUM| |
|
|-------------|--:|--:|--:|--:| |
|
| dev |200|200|200|200| |
|
| test |200|200|200|200| |
|
|
|
## Additional Information |
|
|
|
### Citation Information |
|
|
|
**BibTeX:** |
|
|
|
````BibTeX |
|
@inproceedings{skandalis-etal-2024-new-datasets, |
|
title = "New Datasets for Automatic Detection of Textual Entailment and of Contradictions between Sentences in {F}rench", |
|
author = "Skandalis, Maximos and |
|
Moot, Richard and |
|
Retor{\'e}, Christian and |
|
Robillard, Simon", |
|
editor = "Calzolari, Nicoletta and |
|
Kan, Min-Yen and |
|
Hoste, Veronique and |
|
Lenci, Alessandro and |
|
Sakti, Sakriani and |
|
Xue, Nianwen", |
|
booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)", |
|
month = may, |
|
year = "2024", |
|
address = "Torino, Italy", |
|
publisher = "ELRA and ICCL", |
|
url = "https://aclanthology.org/2024.lrec-main.1065", |
|
pages = "12173--12186", |
|
abstract = "This paper introduces DACCORD, an original dataset in French for automatic detection of contradictions between sentences. It also presents new, manually translated versions of two datasets, namely the well known dataset RTE3 and the recent dataset GQNLI, from English to French, for the task of natural language inference / recognising textual entailment, which is a sentence-pair classification task. These datasets help increase the admittedly limited number of datasets in French available for these tasks. DACCORD consists of 1034 pairs of sentences and is the first dataset exclusively dedicated to this task and covering among others the topic of the Russian invasion in Ukraine. RTE3-FR contains 800 examples for each of its validation and test subsets, while GQNLI-FR is composed of 300 pairs of sentences and focuses specifically on the use of generalised quantifiers. Our experiments on these datasets show that they are more challenging than the two already existing datasets for the mainstream NLI task in French (XNLI, FraCaS). For languages other than English, most deep learning models for NLI tasks currently have only XNLI available as a training set. Additional datasets, such as ours for French, could permit different training and evaluation strategies, producing more robust results and reducing the inevitable biases present in any single dataset.", |
|
} |
|
|
|
@inproceedings{giampiccolo-etal-2007-third, |
|
title = "The Third {PASCAL} Recognizing Textual Entailment Challenge", |
|
author = "Giampiccolo, Danilo and |
|
Magnini, Bernardo and |
|
Dagan, Ido and |
|
Dolan, Bill", |
|
booktitle = "Proceedings of the {ACL}-{PASCAL} Workshop on Textual Entailment and Paraphrasing", |
|
month = jun, |
|
year = "2007", |
|
address = "Prague", |
|
publisher = "Association for Computational Linguistics", |
|
url = "https://aclanthology.org/W07-1401", |
|
pages = "1--9", |
|
} |
|
```` |
|
|
|
**ACL:** |
|
|
|
Maximos Skandalis, Richard Moot, Christian Retoré, and Simon Robillard. 2024. [New Datasets for Automatic Detection of Textual Entailment and of Contradictions between Sentences in French](https://aclanthology.org/2024.lrec-main.1065). In *Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)*, pages 12173–12186, Torino, Italy. ELRA and ICCL. |
|
|
|
And |
|
|
|
Danilo Giampiccolo, Bernardo Magnini, Ido Dagan, and Bill Dolan. 2007. [The Third PASCAL Recognizing Textual Entailment Challenge](https://aclanthology.org/W07-1401). In *Proceedings of the ACL-PASCAL Workshop on Textual Entailment and Paraphrasing*, pages 1–9, Prague. Association for Computational Linguistics. |
|
|
|
### Acknowledgements |
|
|
|
This work was supported by the Defence Innovation Agency (AID) of the Directorate General of Armament (DGA) of the French Ministry of Armed Forces, and by the ICO, _Institut Cybersécurité Occitanie_, funded by Région Occitanie, France. |