RAG ARCHITECTURE AND AGENTS

Abstract

This article proposes a bibliographical survey to provide a comprehensive overview of the RAG (Retrieval-Augmented Generation) Architecture and the application of agents in the current context of artificial intelligence (AI). The RAG architecture combines information retrieval techniques with generation models, providing a hybrid approach that improves the efficiency and accuracy of responses generated by AI systems. Proper architecture planning is crucial to the success of RAG-based systems, as it directly affects the agents' ability to handle large volumes of data and provide relevant answers. In this article, the fundamental principles of the RAG Architecture are explored, as well as its practical applications, advantages, disadvantages, and implementation challenges. In addition, a critical analysis of the contributions and emerging trends in the literature on the subject was carried out. The results of the research indicate that, although the RAG Architecture offers significant advances in the interaction between agents and data, there are still important challenges to be overcome for its large-scale adoption. This study seeks to contribute to understanding these challenges and provide insights for future research and implementation.