A Paradigm Shift in Adaptive Education
EduVision ITS represents a novel approach to personalized learning, moving beyond static rule-based systems to a fully dynamic, cognitive architecture. By synthesizing Bayesian Knowledge Tracing (BKT), Spaced Repetition Systems (SRS), and Large Language Models (LLMs), we create a digital tutor that truly understands the learner.
Cognitive Modeling (BKT)
Utilizes Bayesian Knowledge Tracing to probabilistically estimate student mastery. The system differentiates between slips, guesses, and true learning, ensuring that the difficulty curve adapts in real-time.
Neuro-Symbolic Architecture
Combines the generative power of LLMs (Llama 3.1) with the structured precision of Knowledge Graphs. This ensures pedagogical accuracy while maintaining natural, empathetic dialogue.
Spaced Repetition (SRS)
Implements the Ebbinghaus Forgetting Curve to schedule reviews at optimal intervals. The system predicts memory decay and reinforces concepts exactly when the student is about to forget them.