Future improvements: real-world user studies, advanced student modeling, and spaced repetition¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The authors recommend a forward-looking agenda for the platform: conduct real-world user studies to assess learning gains, user experience, and the practical utility of generated courses; improve personalization through advanced student modeling such as adding affective state detection; and integrate mechanisms such as spaced repetition. They also plan to explore different LLMs and RAG architectures, adaptability to other STEM domains, accessibility, and explainability.

Design Implications¶

Context¶

Requirements¶

• Real-world user studies to validate learning gains before pedagogical effectiveness claims

Constraints¶

• Current personalization operates on a basic set of attributes; system performance depends on underlying LLM capabilities and possible biases

Target Learners¶

• mathematics students, with planned extension to other STEM domains

Target Learning Goals¶

• measurable learning gains
• improved personalization through affective and advanced student modeling
• long-term retention via spaced repetition

Related Strategies¶

• Multi Agent Ai Math Tutor Architecture
• Call for research clarifying the importance of specific choice-policy elements and assessing a wider range of student outcomes
• Future research agenda on preschool quality: access, younger ages, home-based settings, and practice-level studies
• Advance micro-credentials through a piloting coalition, a rigorous research agenda, and dissemination of best practices
• Broaden instruments, methods, and expert judgment in psychometric LLM evaluation research
• Retrain and evaluate the wellness system on locally collected, clinically validated data, then run a formal user evaluation before wider deployment
• Prioritize real-world testing of AI tools in university settings, environmental impact measurement, and integrated ethical-sustainability policies
• Planned enhancements: multi-turn conversational follow-up, RAG-based personalization, and feedback analytics for AI-assisted instruction
• Best practices for RL in education: evaluate live with actual users at scale, look beyond STEM, and target under-served learner groups including children
• Collaborative cross-cultural research agenda for validating AI-enabled pedagogical accompaniment
• Planned extensions: expanded user studies, expert-in-the-loop RLHF, hybrid retrieval, and hand-tracking-based assessment

Examples¶

-

Key Sources¶

• Lb0