Everyone Deserves a Teacher They Like: Designing Personality-Matched Multimodal Educational Agents in Virtual Reality

Yuchao Zhuo, Yukun Zhao, Xuanyu Wang, Tristan Braud

Abstract: Virtual Reality (VR) provides immersive and interactive learning environments in which virtual humans can act as educational agents, supporting instruction through embodied interaction. Prior research suggests that learners respond to such agents as social actors and that personality plays an important role in shaping motivation and learning experiences. However, existing studies on personality-adaptive educational agents often focus on isolated personality traits of the Five-Factor Model (FFM), rely on single-modality expression, and lack automatic mechanisms for matching agent personalities to individual learners. To address these limitations, we propose a multimodal educational agent system that automatically matches agent personalities to individual learners based on empirical data from user studies. The system expresses different personalities through verbal and nonverbal behaviors, enabling more engaging interactions in learning environments. Our work explores the potential of personality-aware recommendation and multimodal expression to enhance perceived engagement and support personalized learning experiences in virtual educational environments.

External IDs:doi:10.1145/3772363.3798875