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

Authors: Triet Minh Thai, Ngan Ha-Thao Chu, Anh Tuan Vo, and Son T. Luu.

Description: The UIT-ViCoV19QA dataset comprises 4,500 Vietnamese question-answer pairs about COVID-19 pandemic collected from trusted medical FAQ sources, each question has at least one answer and at most four unique paraphrased answers.

The statistics of the dataset are shown in the table below.

No.	Stats.	Train	Dev.	Test	All
Answer 1	Number of question-answer pairs	3500	500	500	4500
	Average question length	31.44	33.66	32.32	31.79
	Average answer length	120.53	116.04	118.11	119.76
	Question vocabulary size	4396	1869	1770	4924
	Answer vocabulary size	8537	3689	3367	9411
Answer 2	Number of question-answer pairs	1390	209	201	1800
	Average question length	35.56	39.22	39.72	36.45
	Average answer length	40.54	39.25	42.73	40.64
	Question vocabulary size	2883	1269	1207	3305
	Answer vocabulary size	2632	1098	1129	2949
Answer 3	Number of question-answer pairs	542	79	79	700
	Average question length	34.77	36.7	39.28	35.49
	Average answer length	28.68	26.43	30.89	28.67
	Question vocabulary size	1836	717	693	2111
	Answer vocabulary size	1554	503	585	1753
Answer 4	Number of question-answer pairs	272	39	39	350
	Average question length	36.57	37.59	42.15	37.1
	Average answer length	29.75	29.03	35.72	30.25
	Question vocabulary size	1315	470	460	1519
	Answer vocabulary size	924	353	374	1075

Link to publication: https://aclanthology.org/2022.paclic-1.88/.

The dataset is used only for research purposes.

Some parts of the source code were inherited from the publication at https://github.com/barshana-banerjee/ParaQA_Experiments.git.

Contact information

Mr. Triet Minh Thai: [email protected]
Mr. Son T. Luu: [email protected]

Citation

@inproceedings{thai-etal-2022-uit,
    title = "{UIT}-{V}i{C}o{V}19{QA}: A Dataset for {COVID}-19 Community-based Question Answering on {V}ietnamese Language",
    author = "Thai, Triet  and Thao-Ha, Ngan Chu  and Vo, Anh  and Luu, Son",
    booktitle = "Proceedings of the 36th Pacific Asia Conference on Language, Information and Computation",
    month = oct,
    year = "2022",
    address = "Manila, Philippines",
    publisher = "De La Salle University",
    url = "https://aclanthology.org/2022.paclic-1.88",
    pages = "801--810",
}