JIITA, vol.9 no.1, p.1039-1050, DOI: 10.22664/ISITA.2025.9.1.1039
In-Hye Park, Min-Jeong Kim, Kyung-Ae Cha
Abstract. Recent advancements in language models have led to their widespread application across various fields, achieving remarkable success. However, these models often exhibit a phenomenon known as ‘hallucination’, where they generate responses that present false information as factual. This issue has emerged as a critical challenge, particularly in scenarios where reliability and accuracy are paramount. To address this problem, the RAG (Retrieval-Augmented Generation) technique has gained significant attention. RAG combines generative language models with information retrieval systems, allowing the model to search for relevant information from external databases before generating a response. This approach helps mitigate hallucinations and delivers more trustworthy and accurate information. This study leverages the LangChain framework to develop a RAG-based Q&A system using university policies and administrative regulations as the dataset, which serves as the foundation for students’ academic and institutional operations. By systematically organizing and managing complex data, the proposed system aims to provide accurate and reliable responses, offering users valuable and actionable information efficiently.
Keywords; Large Language Models (LLMs), Retrieval-Augmented Generation, LangChain
Fullpaper: