The science behind real learning
Not quiz-app guesswork. Research-grade adaptive education.
Bayesian Knowledge Tracing (BKT)
BKT is a probabilistic model originally developed at Carnegie Mellon University. It tracks the probability that a student has mastered a specific concept based on their answer history — not just right/wrong streaks, but what each answer reveals about underlying understanding. Aauti Learn runs BKT independently for each of 976+ curriculum concepts.
Real-Time Difficulty Adjustment
Every question is calibrated to the student's current mastery probability. Too easy? Questions escalate. Struggling? The system steps back to prerequisite concepts. The target is the Vygotsky Zone of Proximal Development — challenging enough to grow, not so hard it's discouraging.
Spaced Repetition
Mastered concepts resurface at optimal review intervals based on forgetting curves. So what Ishan learned in October is still there in January — because the system scheduled review at exactly the right moment.
14 Teaching Personas
Pixel adapts teaching style based on student emotional state, recent performance, and time-of-day. Struggling after school? Pixel switches to encouragement mode with simpler examples. Confident and fast? Pixel pushes harder with extension challenges.
Socratic Hints — Not Answers
When a student asks for help, Pixel never just gives the answer. Instead, Pixel asks a leading question that gets the student to reason it out themselves. Learning happens in the reasoning, not the answer.
Parent Transparency
Parents see concept-level mastery, not just time-on-task. The dashboard shows exactly which fraction types a child struggles with — not just "30 minutes of Math." Weekly email reports summarize what was mastered, what needs more work, and recommended next sessions.