Escape from dungeon—modeling user intentions with natural language processing techniques

Abstract

Educational games are a powerful solution for pedagogical problems, both from students’ and teachers’ points of view. Students may experience the inability to focus on lectures in a traditional learning environment, as they cannot understand the lecture materials, are not motivated to study, or the subjects are not challenging enough. Research on learning strategies shows that students are more likely to remain focused and engaged in a smart learning environment that makes use of gamification, instead of a classical classroom scenario, where teachers present formal lectures. Our game, Escape from Dungeon, falls in the category of serious games for problem solving that integrate natural language processing (NLP) techniques adopted to model user intentions. We focused on ensuring appealing graphics and ease of interaction, while relying on novel technologies. The main character of the game is controlled through vocal commands that are interpreted using NLP tools. The game was tested by ten users throughout a pilot test. Users considered the game innovative and entertaining. However, users suggested additional game scenes for an extended gameplay, as well as more actions and intents to be covered within the interaction with the character.