metadata

license: cc-by-nc-4.0
task_categories:
  - question-answering
language:
  - az
tags:
  - NLP
  - Azerbaijani NLP
  - AzLLM
  - AzNLP
  - QA
  - Question Answering
pretty_name: >-
  Question-Answering Dataset for Azerbaijani Language based on Intellectual
  Games (3sual.az)
size_categories:
  - 1K<n<10K

Question-Answering Dataset for Azerbaijani Language based on Intellectual Games (3sual.az)

Baku Higher Oil School Research and Development Center on AI introduces a dataset to fine-tune the NLP models to manage it as a question answering. This dataset contains 4697 questions with answers and explanations. In some cases answer does not exist therefore that slot is empty. Dataset have been collected from 3sual.az and copyright belongs to corresponding website (3sual.az) and its owner Bahruz Aydinli. Only portion of the questions have been collected while it is also considered that the names of the authors exist per question.

This public source intented to be used for research purposes and prohibited to use for commercial.

Dataset contains these fields:

Question: Question given in the website

Answer: Short answer for corresponding question

Explanation: Brief explanation why answer is correct for the given question

Author: Author of the question.

Small considerations have to be taken:

Explanation: None - Means that there is no wide explanation of the answer.
Some questions is intended to have image also. They are not separated.
When there is only explanation but no answer, answer is assigned as None.

Dataset have been published by Baku Higher Oil School Research and Development Center on AI. This R&D Center has aim to develop ecosystem of the Azerbaijani NLP in order to accerate the development in all firms.