metadata

languages:
  - ca

VilaQuAD, An extractive QA dataset for catalan, from Vilaweb newswire text

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.4562337

Introduction

This dataset contains 2095 of Catalan language news articles along with 1 to 5 questions referring to each fragment (or context). VilaQuad articles are extracted from the daily Vilaweb (www.vilaweb.cat) and used under CC-by-nc-sa-nd (https://creativecommons.org/licenses/by-nc-nd/3.0/deed.ca) licence. This dataset can be used to build extractive-QA and Language Models.

Supported Tasks and Leaderboards

Extractive-QA, Language Model

Languages

CA- Catalan

Directory structure

README.md
dev.json
test.json
train.json
vilaquad.py

Dataset Structure

Data Instances

Three json files

Data Fields

Follows ((Rajpurkar, Pranav et al., 2016) for squad v1 datasets. (see below for full reference)

Example:

{
  "data": [
    {
      "title": "Com celebrar el Cap d'Any 2020? Deu propostes per a acomiadar-se del 2019",
      "paragraphs": [
        {
          "context": "Hi ha moltes propostes per a acomiadar-se d'aquest 2019. Els uns es queden a casa, els altres volen anar lluny o sortir al teatre. També s'organitzen festes o festivals a l'engròs, fins i tot hi ha propostes diürnes. Tot és possible per Cap d'Any. Encara no sabeu com celebrar l'entrada el 2020? Us oferim una llista amb deu propostes variades arreu dels Països Catalans: Festivern El Festivern enguany celebra quinze anys.",
          "qas": [
            
            {
              "answers": [
                {
                  "text": "festes o festivals",
                  "answer_start": 150
                }
              ],
              "id": "P_23_C_23_Q2",
              "question": "Què s'organitza a l'engròs per acomiadar el 2019?"
            },
            ...
          ]
        }
      ]
    }, 
    ...
   ]
}

Data Splits

train.json: 1295 contexts, 3882 questions dev.json: 400 contexts, 1200 questions test.json: 400 contexts, 1200 questions

Content analysis

Number of articles, paragraphs and questions

Number of contexts: 2095
Number of questions: 6282
Questions/context: 2.99
Number of sentences in contexts: 11901
Sentences/context: 5.6

Number of tokens

tokens in context: 422477
tokens/context 201.66
tokens in questons: 65849
tokens/questions: 10.48
tokens in answers: 27716
tokens/answers: 4.41

Question type

Question	Count	%
què	1698	27.03 %
qui	1161	18.48 %
com	574	9.14 %
quan	468	7.45 %
on	559	8.9 %
quant	601	9.57 %
quin	1301	20.87 %
no question mark	0	0.0 %

Question-answer relationships

From 100 randomly selected samples:

Lexical variation: 32.0%
World knowledge: 16.0%
Syntactic variation: 22.0%
Multiple sentence: 16.0%

Dataset Creation

Methodology

From a the online edition of the catalan newspaper Vilaweb (https://www.vilaweb.cat), 2095 articles were randomnly selected. These headlines were also used to create a Textual Entailment dataset. For the extractive QA dataset, creation of between 1 and 5 questions for each news context was commissioned, 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, 6282 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. We also created another QA dataset with wikipedia to ensure thematic and stylistic variety.

Source Data

https://www.vilaweb.cat/

Initial Data Collection and Normalization

The source data are scraped articles from archives of Catalan newspaper website Vilaweb (https://www.vilaweb.cat).

Who are the source language producers?

[More Information Needed]

Annotations

Annotation process

We comissioned 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?

Annotation was commissioned to an specialized company that hired a team of 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.