Dipartimento di Ingegneria informatica, automatica e gestionale

In recent years, there has been an exponential growth in the demand for distance learning. In addition, the pandemic from COVID-19 has radically changed the techniques but also the teaching/ learning methodologies. In such a picture, Massive Online Open Courses (MOOCs), are increasingly emerging on the Web. Due to the big number of students, in MOOCs the traditional monitoring and instructional interventions, performed by the teacher on individual learners, are of difficult, if not impossible, application. An Artificial Intelligence approach, based on Virtual Conversational Agents or Intelligent Chatbots, can help overcoming such difficulties. In this paper we present a system, at its early stage of development and based mainly on a deep learning model, able, at different levels, to suggest didactic material to students in a query/answer modality. It is able to answer students’ questions by proposing didactic material taken both from a specific knowledge domain or from Wikipedia. We investigate the potential of such an early stage implementation, through several case studies. We also present a qualitative evaluation, based on the case studies findings, which we think is encouraging towards the development and field experimentation of a whole system.