Teaching AI algorithms using animations reinforced by interactive exercises

Abstract

This paper describes our experience using laboratories in the teaching of an undergraduate subject in artificial intelligence (AI). The presentation of key AI algorithms in lectures was replaced by a highly structured set of exercises undertaken by students in supervised laboratory sessions. The exercises, which were the students’ first contact with the algorithms, used a graphic animation of the algorithm, followed by active problem solving using a computer implementation of the same algorithm. As an environment for introducing students to new material, this laboratorybased approach encouraged the students to engage deeply with the material from the start. The sessions were very popular as evidenced by responses from student surveys, and anecdotal evidence suggests that the material was learned better. We suggest that a similar approach may be effective in other areas of computer science and in other disciplines.