Supporting Conceptual Awareness with Pedagogical Agents

Abstract

This paper describes a series of efforts in building and conceptualizing software agents for distributed collaborative learning. The agents are referred to as pedagogical agents. We have integrated pedagogical agents within two collaborative environments, TeamWave Workplace and Future Learning Environment. The role of agents in these environments differs from past work on software agents in their function as extended awareness mechanisms, focusing on task and concept awareness (conceptual awareness). Our approach is stimulated by Mead's theory of the 'Generalized Other'. The agents collect statistical information on user activity and analyze the information based on principles of collaboration and knowledge building (participation, group interaction, and scientific discourse). Furthermore, the agents define a trajectory in a pedagogical agent design space, which we define in terms of four dimensions: presentation, intervention, task, and pedagogy. We end the paper by comparing our approach with related work.

Computer-Supported Collaborative Learning: Basic Concepts, Multiple Perspectives, and Emerging Trends

“…Another debate is to what extent automated feedback should be proactive, reactive or latent (Mørch et al, 2005). A system that provides sentence openers and step-by-step guidance is proactive.…”

“…If it allows "wrong" actions to be taken before it gives hints and critique, it is reactive. If the system does not take any suggestive action on its own at all, but allows the learner to request guidance upon demand, it is latent (Mørch et al, 2005). All three intervention strategies are important in CSCL environments, but not at the same time.…”

SAMOS: a model for monitoring students' and groups' activities in collaborative e-learning

“…Groups which were supported with feedback were also more proactive in the planning of their group process. Morch et al (2005) focused on how groups of dispersed people are enabled to work and learn together through enriched web-based environments which are capable of capturing important information from group interaction, but so far they fail to provide a useful and easily handled representation of this information, which may have a complex structure.…”

A Method to Automate the Prediction of Student Academic Performance from Early Stages of the Course

“…Meanwhile, Paolucci et al [8] focus interventions in guiding students to solve academic problems. According to the criteria used by [9], it is possible to intervene in the work of students not only to optimize solutions but also to ensure optimal collaboration between classmates. Some alternatives have already been explored to include artificial intelligence (in some of its variants) to the field of teaching, such as incorporating virtual assistants adapted to specific classes and different levels of students [10] or in online courses to mitigate the impact of the volume of students [11].…”