Key points are not available for this paper at this time.
This paper presents AIA-PAL (Artificial Intelligence Agents for Personalized Adaptive Learning), a framework designed to improve student outcomes through advanced Human-Agent Interaction (HAI). It addresses limitations in current intelligent tutoring systems by enabling dynamic scafolding and personalized learning paths. AIA-PAL uses LangGraph for decision-making and a multi-agent system via CrewAI to adapt learning in real time. Its dual-layer approach integrates real-time monitoring with specialized agents (tutor, practical, and teacher) to provide targeted pedagogical support. Large language models are employed to enable personalized dialogue, improving responsiveness to queries. To ensure pedagogical accuracy and minimize hallucinations, AIA-PAL uses retrieval-augmented generation grounded in teacher-provided content and validated through structured human oversight. Preliminary results indicate improved scafolding and personalized learning, boosting engagement and efficiency. The paper details the framework’s architecture, methods, and outcomes, highlighting real-time adaptability and HAI as key to effective learning.
Hedi et al. (Wed,) studied this question.