DIGITAL LIBRARY
SIMULATION OF CONCURRING MULTI-AGENT EVENTS WITH INDIVIDUALIZED EPISTEME TO MAXIMIZE VARIABILITY OF PEDAGOGICAL SITUATIONS FOR IMMERSED INTELLIGENT TUTORING SYSTEM
Kyungpook National University (KOREA, REPUBLIC OF)
About this paper:
Appears in: ICERI2018 Proceedings
Publication year: 2018
Pages: 10013-10018
ISBN: 978-84-09-05948-5
ISSN: 2340-1095
doi: 10.21125/iceri.2018.0871
Conference name: 11th annual International Conference of Education, Research and Innovation
Dates: 12-14 November, 2018
Location: Seville, Spain
Abstract:
In contrast to conventional narrative-based systems mainly pursuing dramatic interests, virtual world-based education systems strive to provide realistic pedagogical experiences in immersed situations. The diversity and fidelity of the situations unfolding in the virtual world determines the quality of situated pedagogical experience. We develop a virtual world simulation system with maximum variability and authenticity of situations for immersed intelligent tutoring systems. Our virtual world simulation model designs a pedagogical situation to generally comprise many concurring events from an overarching multi-agent event, or independent events to be coincidentally inter-coupled into an emergent situation. Our model employs dynamic coupling among events in contrast to the conventional narrative-based systems like Interactive Storytelling, where events are pre-authored in a monolithic manner along the fixed storylines. This dynamic approach allows numerous multi-event situations to be generated by instantiating the schematic events with parameters corresponding to its associated background-world conditions of each agent. This already-complex situation would be further compounded by their associated agents’ individualized episteme, i.e., their respective belief on the real world, instead of an omniscient view as in the conventional simulated world-based systems like Edu-games. Individualized episteme is designed to be incessantly updated according to renewed background-world conditions to adapt their behavior proactively and reactively. Still such episteme cannot be perfect due to incomplete information and partial observation with respect to the situations they are involved. This leads to their misconception and epistemic discrepancies between the up-to-second world conditions and the agent’s episteme, further deepening the situation variability. The resulting enhanced diversity and fidelity of these virtual-world situations greatly improves the quality of immersed pedagogical experience in intelligent tutoring systems based on our virtual world simulation. To support this simulation method in its implementation, a real-time execution scheme is devised to naturally couple independent events into a situation in the fine level of agents’ actions. Our simulation method affords the highest diversity and fidelity of complex situations involving multiple agents, still achieves much enhanced authoring scalability over monolithically-authored conventional simulation systems.
Keywords:
Immersive pedagogical experience, intelligent tutoring system, virtual world simulation, individualized episteme.