Chatbot conversacional inteligente basado en la Comprensión del lenguaje natural y modelo de aprendizaje auto-regresivo GPT-3 para admisiones de la Universidad

An intelligent conversational chatbot using state-of-the-art technologies in natural language processing and machine learning, based on the GPT-3 autoregressive language model, is presented. The ability of the GPT-3 model to generate coherent and fluent text is highlighted, and some of its applicati...

Full description

Saved in:
Bibliographic Details
Published inRISTI : Revista Ibérica de Sistemas e Tecnologias de Informação no. E66; pp. 1 - 12
Main Authors Quisi-Peralta, Diego F, Lopez-Nores, Martin, Salgado, Juan P, Murillo-Valarazo, Jordan, Garcia-Velez, Roberto, Galan-Mena, Jorge
Format Journal Article
LanguageEnglish
Published Lousada Associação Ibérica de Sistemas e Tecnologias de Informacao 01.02.2024
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:An intelligent conversational chatbot using state-of-the-art technologies in natural language processing and machine learning, based on the GPT-3 autoregressive language model, is presented. The ability of the GPT-3 model to generate coherent and fluent text is highlighted, and some of its applications are described, in this context the use within a chatbot for the admissions area within the university. The functional requirements, development, implementation, and evaluation of the proposed solution are presented, showing promising results in terms of the chatbot's effectiveness in understanding natural language and the ability to maintain a useful conversation with applicants with a 90% satisfaction rate. En el presente artículo, se presenta un chatbot conversacional inteligente que combina tecnologías de vanguardia en el campo del procesamiento del lenguaje natural y el aprendizaje automático, basado en el modelo de lenguaje auto-regresivo GPT-3 que es un modelo de lenguaje generativo basado en redes neuronales que puede generar texto coherente y fluido (Zhang, M., & Li, J., 2021).
ISSN:1646-9895