Datasets:
languages:
- ca
ViquiQuAD, An extractive QA dataset for catalan, from the Wikipedia
BibTeX citation
If you use any of these resources (datasets or models) in your work, please cite our latest paper:
@inproceedings{armengol-estape-etal-2021-multilingual,
title = "Are Multilingual Models the Best Choice for Moderately Under-resourced Languages? {A} Comprehensive Assessment for {C}atalan",
author = "Armengol-Estap{\'e}, Jordi and
Carrino, Casimiro Pio and
Rodriguez-Penagos, Carlos and
de Gibert Bonet, Ona and
Armentano-Oller, Carme and
Gonzalez-Agirre, Aitor and
Melero, Maite and
Villegas, Marta",
booktitle = "Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021",
month = aug,
year = "2021",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2021.findings-acl.437",
doi = "10.18653/v1/2021.findings-acl.437",
pages = "4933--4946",
}
Digital Object Identifier (DOI) and access to dataset files
https://doi.org/10.5281/zenodo.4562345
Introduction
This dataset contains 3111 contexts extracted from a set of 597 high quality original (no translations) articles in the Catalan Wikipedia "Viquipèdia" (ca.wikipedia.org), and 1 to 5 questions with their answer for each fragment.
Viquipedia articles are used under [CC-by-sa] (https://creativecommons.org/licenses/by-sa/3.0/legalcode) licence.
This dataset can be used to fine-tune and evaluate extractive-QA and Language Models. It is part of the Catalan Language Understanding Benchmark (CLUB) as presented in:
Armengol-Estapé J., Carrino CP., Rodriguez-Penagos C., de Gibert Bonet O., Armentano-Oller C., Gonzalez-Agirre A., Melero M. and Villegas M.,Are Multilingual Models the Best Choice for Moderately Under-resourced Languages? A Comprehensive Assessment for Catalan". Findings of ACL 2021 (ACL-IJCNLP 2021).
Supported Tasks and Leaderboards
Extractive-QA, Language Model
Languages
CA- Catalan
Directory structure
- README
- dev.json
- test.json
- train.json
- viquiquad.py
Dataset Structure
Data Instances
json files
Data Fields
Follows ((Rajpurkar, Pranav et al., 2016) for squad v1 datasets. (see below for full reference)
Example:
{ "data": [ { "title": "Frederick W. Mote", "paragraphs": [ { "context": "L'historiador Frederick W. Mote va escriure que l'ús del terme \\\\\\\\\\\\\\\\"classes socials\\\\\\\\\\\\\\\\" per a aquest sistema era enganyós i que la posició de les persones dins del sistema de quatre classes no era una indicació del seu poder social i riquesa reals, sinó que només implicava \\\\\\\\\\\\\\\\"graus de privilegi\\\\\\\\\\\\\\\\" als quals tenien dret institucionalment i legalment, de manera que la posició d'una persona dins de les classes no era una garantia de la seva posició, ja que hi havia xinesos rics i amb bona reputació social, però alhora hi havia menys mongols i semu rics que mongols i semu que vivien en la pobresa i eren maltractats.", "qas": [ { "answers": [ { "text": "Frederick W. Mote", "answer_start": 14 } ], "id": "5728848cff5b5019007da298", "question": "Qui creia que el sistema de classes socials de Yuan no s’hauria d’anomenar classes socials?" }, ... ] } ] }, ... ] }
Data Splits
train.development,test
Content analysis
Number of articles, paragraphs and questions
- Number of articles: 597
- Number of contexts: 3111
- Number of questions: 15153
- Questions/context: 4.87
- Number of sentences in contexts: 15100
- Sentences/context: 4.85
Number of tokens
- tokens in context: 469335
- tokens/context 150.86
- tokens in questions: 145249
- tokens/questions: 9.58
- tokens in answers: 63246
- tokens/answers: 4.17
Lexical variation
After filtering (tokenization, stopwords, punctuation, case), 83,88% of the words in the question can be found in the Context
Question type
Question | Count | % |
---|---|---|
què | 4220 | 27.85 % |
qui | 2239 | 14.78 % |
com | 1964 | 12.96 % |
quan | 1133 | 7.48 % |
on | 1580 | 10.43 % |
quant | 925 | 6.1 % |
quin | 3399 | 22.43 % |
no question mark | 21 | 0.14 % |
Question-answer relationships
From 100 randomly selected samples:
- Lexical variation: 33.0%
- World knowledge: 16.0%
- Syntactic variation: 35.0%
- Multiple sentence: 17.0%
Dataset Creation
Methodology
From a set of high quality, non-translation, articles in the Catalan Wikipedia (ca.wikipedia.org), 597 were randomly chosen, and from them 3111, 5-8 sentence contexts were extracted. We commissioned creation of between 1 and 5 questions for each context, following an adaptation of the guidelines from SQUAD 1.0 [Rajpurkar, Pranav et al. “SQuAD: 100, 000+ Questions for Machine Comprehension of Text.” EMNLP (2016)], (http://arxiv.org/abs/1606.05250). In total, 15153 pairs of a question and an extracted fragment that contains the answer were created.
Curation Rationale
For compatibility with similar datasets in other languages, we followed as close as possible existing curation guidelines.
Source Data
Initial Data Collection and Normalization
The source data are scraped articles from the Catalan wikipedia site (https://ca.wikipedia.org).
Who are the source language producers?
[More Information Needed]
Annotations
Annotation process
We commissioned the creation of 1 to 5 questions for each context, following an adaptation of the guidelines from SQUAD 1.0 (Rajpurkar, Pranav et al. “SQuAD: 100, 000+ Questions for Machine Comprehension of Text.” EMNLP (2016)), http://arxiv.org/abs/1606.05250.
Who are the annotators?
Native language speakers.
Dataset Curators
Carlos Rodríguez and Carme Armentano, from BSC-CNS
Personal and Sensitive Information
No personal or sensitive information included.
Considerations for Using the Data
Social Impact of Dataset
[More Information Needed]
Discussion of Biases
[More Information Needed]
Other Known Limitations
[More Information Needed]
Contact
Carlos Rodríguez-Penagos ([email protected]) and Carme Armentano-Oller ([email protected])
License
This work is licensed under a Attribution-ShareAlike 4.0 International License.