7 / 2020-01-28 01:57:30

Virtual Assistant Integrated with AVA FURG as Academic Support Resource: an experience with recurrent neural networks

Virtual assistant,IBM Watson,Virtual learning environment,Recurring Neural Networks

This Full paper research presents an intelligent chatbot, developed through the IBM cognitive services platform, to improve the attendance to students of higher education courses in the use of the Virtual Learning Environment (VLE). Recurrent Neural Networks (RNR) were used to generate a model to classify the possible intentions of users. 48 students interacted with the virtual assistant and, at the end, answered a satisfaction questionnaire that gave the chatbot a high degree of satisfaction. The trained artificial neural network was submitted to the main evaluation criteria and obtained 99.80% in the accuracy tests. For the classification of intentions, there was a significant acceptance of the chatbot by the students. The virtual assistant also demonstrated the ability to collaborate in student service activities in the frequent doubts regarding the use of VLE, since it can be accessed through a computer or a mobile device. This study presented evidence that demonstrated that it is possible and viable to implant the chatbot to assist in the context of learning.