MATCHFAI: Sistema de recomendação híbrido de músicas usando inteligência artificial generativa e agentes inteligentes, 2024

Abstract

This work explores the use of Large Language Models (LLMs) in a hybrid music recommendation system, combined with artificial intelligence agents, presenting a new approach to music recommendation. The proposed architecture includes various agents to perform hybrid filtering, combining content-based and collaborative filtering, along with other support agents. Several LLMs, such as Gemini, LLaMA, and Mixtral, are employed as recommendation engines. The final results show that the Gemini 1.5 Pro model achieved the best performance among the evaluated models, based on data obtained from user testing.