The Optimal Learning Lab studies how memory models, knowledge tracing, and adaptive educational systems can be turned into practical tools for instruction. Our work connects theory, data, and software design so that learning environments can make better decisions about timing, sequence, and support.
SPARC combines learner models, explicit instructional rules, and generative AI to study adaptive tutoring. TutorScript is the language used to express those rules: it makes the choice of instructional action available for inspection, while AI generates the selected explanation, example, figure, or question. These emerging research capabilities are being tested in controlled deployments.
Chu, Pavlik, and Hu's How meaningfulness and initial learning degree influence associative storage and retrieval processes: A multinomial processing tree and power function analysis was published April 15, 2026, in Psychonomic Bulletin & Review, 33, article 141. Read the published article.
We build and evaluate formal models of learning and memory that capture spacing, recency, practice history, and changing performance.
We translate those models into educational software that schedules practice, prioritizes items, and adapts to the learner's history.
We investigate learning beyond isolated facts, including connected concepts, dependencies across ideas, and instructional structure.
Better learning systems should not just deliver content. They should make defensible decisions about when, what, and how practice should happen.
That idea sits at the center of the lab's work across memory, modeling, and instructional software.
