A Conversational Agent for the Improvement of Human Reasoning Skills

Abstract

Human reasoning is the ability to make sound and goal-oriented decisions and is therefore highly relevant in daily life. However, its importance has not yet been addressed explicitly in education. In this paper, we propose to develop a computer-based conversational agent named Liza for improving human reasoning skills. Liza is able to hold conversations with humans to help them solve a small selection of well-studied reasoning problems. Such a conversational agent allows a much more scalable and inexpensive approach for teaching at least basic reasoning principles. The evaluation study shows that the conversational agent Liza improved the reasoning skills of the participants, who had conversations with the agent to solve reasoning problems and that the group using Liza achieved much better learning effects than a group studying with a non-interactive online course that was implemented as a control condition.