Estudos de algoritmos de aprendizagem profunda no contexto de processamento de linguagem natural para desenvolvimento de assistentes virtuais

Abstract

Teaching is an area that has many challenges, we can mention some of them as the quality and the task of solving students’ doubts are some problems that are found every day in Brazilian schools and universities. Every year the Natural Language Processing (NLP) models gain more power. The variety of applicability is huge, for example a question and answer system, which is a system capable of receiving an answer and taking into account a database, has the ability to return the best expected answer. It is with this in mind that we did a study on BERT, which is a deep learning algorithm (Deep Learning) from Google, and we sought to understand its functioning, applicability, feasibility and its results in the development of virtual assistants and question and answer systems. for use in the educational environment.