Projects
Our research of digital game-based learning focuses on using an integrative approach to design and examine scalable game and participatory simulation based learning systems that comprise learning-oriented game mechanics, design-based pedagogy, data-driven knowledge tracing, and adaptive learner support. Current projects include: "E-Rebuild" - Mathematical Thinking and Learning via Architectural Design and Modeling (Ke as PI, Funded by National Science Foundation, ; ), Virtual-Reality-Based Social and Cognitive Skills Training for Children with High Functioning Autism (Ke as PI, Funded by , Grant #201400178; ), and Game-based Assessment and Support of STEM-related Competencies and Interest (Ke as Co-PI, Funded by ).
We design and study inclusive and immersive e-learning environments that promote engaging and effective learning interactions for a diversified learner population. The project of MILE (Mixed-Reality-Integrated Learning Environment) (Ke as PI, Funded by ) aims to study the design model and effects of an AI-infused, virtual world-based training platform to provide responsive teaching practice to preservice teachers and STEM graduate teaching assistants. Integrating 3D virtual reality, LLM-driven virtual human, and body sensory technology, this training platform for teaching will enable student instructors to practice, observe, and reflect on teaching in a variety of instructional settings, and it will provide them with a deep understanding of learning and teaching strategies through active experimentation and problem solving. This project builds on an earlier project of (Ke as PI, Funded by , Grant #200800124) that utilizes qualitative and quantitative techniques to explore the impact on learning and success of online pedagogies and contexts for students living in rural and urban areas.
Building upon the MILE platform, Evelyn project extends this work into a real-time 3D classroom simulation for preservice teacher training. Preservice teachers navigate an immersive virtual classroom as embodied avatars, engaging with LLM-driven virtual student agents that generate dynamic, context-sensitive dialogue in response to teacher input. At the core of this iteration is a formal cognitive-affective state dependency modela bidirectional, weighted, directed graph of constructs including arousal, sense of mastery, self-efficacy, and metacognitive awarenessthat governs how each virtual students cognitive and emotional states evolve across the interaction. The platform supports administrator oversight and a structured lesson-and-task system, creating a scalable training environment where preservice teachers can practice responsive teaching and directly observe the downstream effects of their instructional decisions on student state. Try our Evelyn !