Assistente conversacional para resolução de Problemas Trigonométricos em Linguagem Natural

Intelligent Tutoring Systems (ITS) are an interdisciplinary area aiming to investigate how to model instructional content based on pedagogical decisions and students’ interactions. Those decisions use student modeling, basedrules domain knowledge and teaching strategies. Most ITS advocate that the feedback of exercises and activities planned by the system define the knowledge students’ level. Therefore, limiting the student’s interactive possibility through the conversational interface. A more “wizard” approach integrates Chatbots into the ITS, using artificial intelligence techniques to manage interactions. This article intends to provide a greater degree of freedom for the Kuyven et al. RBIE V.28 – 2020 209 student through natural language inputs of problems out of the activity set of the ITS. This work presents a Recurrent Neural Network (RNR) model, which translates trigonometric problems, entered by students, into equation models as part of a larger ITS Trigonometry project. The experiments conducted showed that the proposed model correctly classified a large part of the problems posed by students, providing answers to the proposed problems in a step-bystep resolution format of a Computer Algebra System (CAS).