A task-oriented dialogue system for moral education

Abstract

We present a novel and practical dialogue system specifically designed for teachers and parents to solve students’ problems in moral education. Guided by the case-based reasoning theory, we collect the high-quality cases and teaching strategies from heterogeneous sources, and then construct the dedicated knowledge graph to manage the large volume of information in this domain. By leveraging on the latest natural language processing techniques, we finally implement a task-oriented dialogue system to precisely understand user’s problem and subsequently recommend possible solutions. We show the great promise of the system for K-12 education and demonstrate how the system solves the problem raised by the teacher for moral education.