Recuperação da informação e inteligência artificial generativa com large language model e retrieval-augmented generation

Abstract

Generative Artificial Intelligence is impacting society as a whole, enabling the development of solutions across many fields of knowledge. In the context of Information Science, techniques such as Large Language Models and Retrieval-Augmented Generation, widely used in tools like ChatGPT and Gemini, have the potential to broadly influence and relate to Information Retrieval. Therefore, this work aims to present an Information Retrieval model using a chatbot as the interface and applying Generative Artificial Intelligence techniques - Large Language Models and Retrieval-Augmented Generation. To achieve this, an applied approach was adopted, along with an exploration of the relevant literature, which led to the construction of the chatbot. As a result, the developed solution, the chatbot, was able to apply Generative Artificial Intelligence techniques while demonstrating how Information Retrieval is present and interconnected with the solution. It is concluded that the development of applied Artificial Intelligence research within the scope of Information Science can contribute to new studies and greater understanding of the use of these emerging technologies in the field.