Learning By Arguing In Argument-Based Machine Learning Framework

We propose an approach for the development of argument-based intelligent tutoring systems in which a domain that can be successfully addressed by supervised machine learning is taught in an interactive learning environment. The system is able to automatically select relevant examples and counter-examples to be explained by the students. The students learn by explaining specific examples, and the system provides automated feedback on students' arguments, including generating hints. The role of an argument-based intelligent tutoring system is then to train the students to find the most relevant arguments. The students learn about the high-level domain concepts and then use them to argue about automatically selected examples. We demonstrate our approach in an online application that allows students to learn through arguments with the goal of improving their understanding of financial statements.