LLMs Based Approach for Quranic Question Answering

Zakia Saadaoui, Zakia Saadaoui, Ghassen Tlig, Fethi Jarray, Fethi Jarray

2024

Abstract

This paper addresses a prominent research gap in Quranic question answering, where current methodologies face challenges in capturing the nuanced aspects of inquiries related to the Quran. By presenting an innovative approach that utilizes Large Language Models (LLMs), including GPt4, Bart, and LLAMA, we aim to overcome these limitations, improving the clarity and precision of responses to Quranic queries. The evaluation of the proposed Quranic Question Answering System, using the F1 score metric, demonstrates encouraging results in comprehending and addressing various Quranic queries. Notably, the application of a chain of thought with Bart achieves an impressive F1-score of 95%. This research offers a distinctive perspective on Quranic question answering through the integration of LLMs.

Paper Citation

in Harvard Style

Saadaoui Z., Tlig G. and Jarray F. (2024). LLMs Based Approach for Quranic Question Answering. In Proceedings of the 20th International Conference on Web Information Systems and Technologies - Volume 1: WEBIST; ISBN 978-989-758-718-4, SciTePress, pages 112-118. DOI: 10.5220/0013012900003825

in Bibtex Style

@conference{webist24,
author={Zakia Saadaoui and Ghassen Tlig and Fethi Jarray},
title={LLMs Based Approach for Quranic Question Answering},
booktitle={Proceedings of the 20th International Conference on Web Information Systems and Technologies - Volume 1: WEBIST},
year={2024},
pages={112-118},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013012900003825},
isbn={978-989-758-718-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 20th International Conference on Web Information Systems and Technologies - Volume 1: WEBIST
TI - LLMs Based Approach for Quranic Question Answering
SN - 978-989-758-718-4
AU - Saadaoui Z.
AU - Tlig G.
AU - Jarray F.
PY - 2024
SP - 112
EP - 118
DO - 10.5220/0013012900003825
PB - SciTePress