FlashLearn

This project presents FlashLearn, an AI-assisted adaptive learning system integrating dynamic knowledge tracing with personalized intervention strategies. Traditional flashcard- based learning systems employ static repetition algorithms that fail to account for individual learning patterns and cognitive load. Our system introduces a three-tiered adaptive framework com- prising cognitive state modeling, dynamic difficulty adjustment, and personalized review scheduling. Through a pre-registered, randomized controlled trial with 150 participants across diverse educational domains (programming, medical terminology, for- eign language), FlashLearn demonstrated statistically significant improvements over traditional systems: knowledge retention increased by 18.3% (95% CI: [12.7%, 24.2%], p < 0.001, Cohen’s d = 0.82), learning efficiency improved by 22.4% (95% CI: [15.8%, 29.3%], p < 0.001, d = 0.71), and cognitive load de- creased by 15.2% (95% CI: [9.8%, 20.9%], p < 0.001, d = 0.64) as measured by NASA-TLX assessments. The system achieved 94.3% test coverage through rigorous AI-assisted development practices. All data, code, and materials are openly available (github.com/flashlearn/study, anonymized OSF repository (link provided to editor/reviewers)).