Teachable Agents in Virtual Learning Environments: a Case Study

As intelligent agent and virtual world technologies advanced to new heights in the current decade, researchers from diverse fields have started to look into possible ways of applying these emerging technologies to improve e-learning systems. Nevertheless, most attempts focused on using pedagogical agents to achieve traditional e-learning goals which place heavy emphasis on learning outcomes. In addition, the high technical barrier presented by the agent oriented software engineering (AOSE) paradigm have limited the complexity and wide acceptance of many pedagogical agents in virtual learning environments (VLE). In this paper, we leverage on our previous research results – the Multi-Agent Development Environment (MADE) which drastically reduced to the technical hurdle of incorporating pedagogical agents into VLEs. We present a VLE augmented with teachable agents designed based on the experiential learning model (ELM) to facilitate the learning of high school science topics and demonstrate the efficacy of MADE in designing complex learning activities, designing teachable agents that complement immersive experiential learning and supporting the design based research (DBR) development model.