MotorIA: Automatic E-Learning Course Generation System

Abstract

Recently, through the availability of open data sources and the existence of advanced techniques of Natural Language Processing (NLP) and Artificial Intelligence, several tools have emerged to support the educational learning process of students and lecturers (experts in different knowledge domains). However, there are very few works which propose tools that automate the e-learning course content development process. Hence, experts in the field of educational training have shown great interest in the emergence of applications that ease this type of process. This paper presents MotorIA, a system which uses NLP and Machine Learning techniques as a means to automate online course content generation from different domains (e.g., computer science, biology, mathematic, etc.). MotorIA is supplied with unstructured information extracted from both Wikipedia services and PDF documents. During the e-learning course content generation process, textual information is pre-processed and structured in a knowledge graph, which must satisfy a set of constraints predefined by experts in the field. In a posterior phase, MotorIA was validated by expert users obtaining good evaluation results.