What is it about?
The paper introduces MufassirQAS, a question-answering system using the Retrieval Augmented Generation (RAG) approach to provide accurate information about Islam. The system addresses the limitations of large language models (LLMs) like ChatGPT, which can generate inaccurate or false information, also known as "hallucination." MufassirQAS draws its answers from a curated database of translated and interpreted Islamic texts, including the Quran, Hadiths (sayings of the Prophet Muhammad), and Islamic catechisms. Here's how it works: ● User Question: A user asks a question related to Islam. ● Vector Search: The system converts the question into a vector representation and searches for similar vectors in its database. ● Context Retrieval: The system retrieves the most relevant chunks of text from the Islamic sources. ● Answer Generation: The retrieved context is combined with the question and fed to an LLM (ChatGPT 3.5 Turbo) to generate an answer. ● Source Citation: MufassirQAS presents the answer along with citations from the specific Islamic texts used.
Featured Image
Photo by Masjid MABA on Unsplash
Why is it important?
● Accurate Islamic Information: Provides a reliable source of information about Islam, minimizing the risk of encountering misinformation. ● Transparency and Trust: Increases user trust by revealing the sources of information and showing how the answer was derived. ● Accessibility: Makes learning about Islam easier for individuals without formal religious education.
Perspectives
● Cultural Sensitivity: Developing a system like this requires careful consideration of cultural and religious sensitivities. The choice of sources and the way answers are presented can significantly impact how the information is received. ● Bias in AI: It's crucial to acknowledge that even with a curated database, biases present in the original texts or introduced during the translation and interpretation process can influence the system's responses.
Dr. Enis Karaarslan
Mugla Sitki Kocman Universitesi
Read the Original
This page is a summary of: A RAG-based Question Answering System Proposal for Understanding Islam: MufassirQAS LLM, SSRN Electronic Journal, January 2024, Elsevier,
DOI: 10.2139/ssrn.4707470.
You can read the full text:
Contributors
The following have contributed to this page
