Desenvolvimento de um Chatbot usando redes de aprendizado profundo

Abstract

Presents the development of a chatbot focused on the UFRN academic community using deep learning, for the purpose of helping students and employees making consultations and answering questions faster and at any given time. A convolutional neural network was used to identify what the user wants, a database was created in order to feed this network, and a series of queries and answers to solve the user's issue was designed. The system was tested on several configurations until one was found to have satisfactory results for every expected input.