Implementing intelligent pedagogical agents in virtual worlds: Tutoring natural science experiments in OpenWonderland

Abstract

Intelligent Pedagogical Agents (IPAs) can be thought of as embodied intelligent agents that are designed for pedagogical purposes to support learning. They can be designed in particular for virtual worlds. Virtual worlds are becoming an interesting medium for engineering education for the properties of visual collaboration abilities providing authentic learning experiences and for the opportunity of providing active learning. However, virtual worlds need more educational support to be more inhabited with increased learning services. Incorporating intelligent pedagogical agents into virtual worlds adds such learning support by adding intelligence, improving believability, and the opportunity to increase communication with an artificial educator. However the implementation of intelligent pedagogical agents and adopting them in a virtual world require several efforts with different aspects of implementation. This paper reports our first prototype implementation of an IPA interacting with a learner and a learning object in natural science experiment in a virtual world while providing supporting multi-modal communication abilities. The IPA has features of text chat based on the Artificial Intelligence Markup Language (AIML), a text-to-speech synthesis function, and non-verbal communication abilities through gesture animation. The implementation is presented through explained scenarios of the IPA tutoring an experiment or monitoring a learner avatar interaction with a learning object in a Virtual World. The IPA & the learning scenarios are implemented in the open source of Open Wonderland.