UIT-ViCoV19QA: A Dataset for COVID-19 Community-based Question Answering on Vietnamese Language

UIT-ViCoV19QA: A Dataset for COVID-19 Community-based Question Answering on Vietnamese Language Triet Minh Thai University of Information Technology VNU-HCM, Vietnam 19522397@gm.uit.edu.vn Ngan Ha-Thao Chu University of Information Technology VNU-HCM, Vietnam 19521882@gm.uit.edu.vn Anh Tuan Vo University of Information Technology VNU-HCM, Vietnam 19521226@gm.uit.edu.vn Son T. Luu University of Information Technology VNU-HCM, Vietnam sonlt@uit.edu.vn Abstract For the last two years, from 2020 to 2021, COVID-19 has broken disease prevention measures in many countries, including Viet- nam, and negatively impacted various aspects of human life and the social community. Be- sides, the misleading information in the com- munity and fake news about the pandemic are also serious situations. Therefore, we present the first Vietnamese community-based ques- tion answering dataset for developing question answering systems for COVID-19 called UIT- ViCoV19QA. The dataset comprises 4,500 question-answer pairs collected from trusted medical sources, with at least one answer and at most four unique paraphrased answers per question. Along with the dataset, we set up various deep learning models as baseline to assess the quality of our dataset and initiate the benchmark results for further research through commonly used metrics such as BLEU, ME- TEOR, and ROUGE-L. We also illustrate the positive effects of having multiple para- phrased answers experimented on these mod- els, especially on Transformer - a dominant ar- chitecture in the field of study. 1 Introduction Community-based Question answering (CQA) is a task of question answering based on the wisdom of the crowd (Zhang et al., 2021). In CQA, informa- tion seekers post their questions on a public website or forum, and other users answer them. This kind of question-answering behavior is popular in peo- ple’s daily basis. For example, Quora1 and Reddit2 are several large forums for question-answering. The CQA enables people to ask and answer questions easily (Qiu and Huang, 2015). Additionally, the de- velopment of question and answering systems means computers can now understand and answer the ques- tions of users. In Vietnam, the year from 2020 to 2021 witnessed the COVID-19 pandemic. Information about the COVID-19 spreading situation, medical care, self- quarantine, vaccination policies, and regulations by the government to prevent the spread of COVID- 19 are essential to citizens. People frequently ask questions about the COVID-19 situation, what to do when contacting COVID-19 patients, the vaccina- tion policies, and more. This is our motivation to construct a dataset to help build a question answer- ing system based on CQA about COVID-19 in Viet- namese. Apart from the dataset, we also propose var- ious baseline models to evaluate our dataset’s qual- ity. This paper has three main contributions summa- rized as follows: 1. We introduce UIT-ViCoV19QA, the first community-based question answering collec- tion about the COVID-19 pandemic for Viet- namese constructed from trusted sources. The dataset comprises 4,500 question-answer pairs and is extended to have up to four unique para- phrased answers per question through an effi- cient paraphrasing process. 2. We assess the dataset’s quality and estab- 1https://www.quora.com/ 2https://www.reddit.com/ lish a future research benchmark through ex- periments with various Sequence-to-Sequence …

Content truncated to 500 words