Case Study on Understanding the Power of Retrieval Augmented Generation (RAG)
Venkata Jaipal Reddy Batthula, Richard S. Segall, Sreejith Sivasubramony
Authors Information |
Citation |
Full Text |
Venkata Jaipal Reddy Batthula
Department of Computer Science, Southwestern College, Winfield, Kansas, United States
Richard S. Segall
Department of Information Systems & Business Analytics, Neil Griffin College of Business, Arkansas State University, State University, Arkansas, United States
Sreejith Sivasubramony
Walmart Inc., Bentonville, Arkansas, United States
Cite this paper as:Batthula, V. J. R., Segall, R. S., Sivasubramony, S. (2025). Case Study on Understanding the Power of Retrieval Augmented Generation (RAG).
Journal of Systemics, Cybernetics and Informatics, 23(6), 27-35. https://doi.org/10.54808/JSCI.23.06.27
Online ISSN (Journal): 1690-4524
Abstract
This paper explores how Generative AI is changing with the use of Retrieval-Augmented Generation (RAG). RAG helps improve Artificial Intelligence (AI) systems by making them more capable, efficient and accurate. The paper explains how to build the Retrieval Augmented Generation model, covering important steps like preparing the data, creating embeddings, and setting up the retrieval system. Through a case study, we look at the main components of RAG, how it works with Large Language Models (LLMs), and why it is important in everyday digital tools. One of the goals is to compare different strategies for RAG, including choices for embeddings, similarity metrics and language models to find an optimal approach that can be generalized to work best. This helps us to understand how these factors affect performance and gives us ideas for building better and more efficient systems.
