WSEAS Transactions on Information Science and Applications
Print ISSN: 1790-0832, E-ISSN: 2224-3402
Volume 21, 2024
AraQA-BERT: Towards an Arabic Question Answering System using Pre-trained BERT Models
Author: Afnan H. Alshehri
Abstract: To increase performance, this study presents AraQA-BERT, an Arabic question-answering (QA) system that makes use of pre-trained BERT models. The study emphasizes how important QA systems are for promptly and accurately responding to user inquiries, especially when those inquiries are made in native tongues. Arabic QA systems are necessary because of the complexity and linguistic variances of the Arabic language, even if English QA systems have made substantial progress. The study examines the use of pre-trained language models, such as AraBERT and Arabic-BERT, for Arabic QA tasks with a focus on Modern Standard Arabic (MSA). The study's contributions include the creation of a web-based application named AraQA-BERT for open-domain QA, trials on TyDi and ARCD datasets, and a methodology for employing pre-trained models.
Search Articles
Keywords: AraQA-BERT, Modern Standard Arabic, Question-answering, QA systems, Linguistic variations, Arabic-BERT, Arabic-BERT
Pages: 361-373
DOI: 10.37394/23209.2024.21.34
WSEAS Transactions on Information Science and Applications, ISSN / E-ISSN: 1790-0832 / 2224-3402, Volume 21, 2024, Art. #34