Apprentice learner models support efficient tutor authoring, more efficient than Example-Tracing with mass production¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2022 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Authoring an experimental design expert model with the Decision Tree model is more efficient than with Example-Tracing using mass production. → Christopher J. MacLellan & Kenneth R. Koedinger 2022

Evidence¶

Christopher J. MacLellan & Kenneth R. Koedinger 2022¶

Christopher J. MacLellan & Kenneth R. Koedinger. (2022). Domain-General Tutor Authoring with Apprentice Learner Models. International Journal of Artificial Intelligence in Education. https://doi.org/10.1007/s40593-020-00214-2

q2 · i? · design · r2

A case study comparing authoring efficiency for an experimental design tutor. The Example-Tracing model required 168 demonstrations and 24 unordered group specifications across 3 behavior graph templates, estimated at 26.96 minutes using a keystroke-level model. The Decision Tree model was trained on 20 problems interactively.

"we estimate that the behavior graphs for the experimental design tutor would take 26.96 minutes to build using Example-Tracing"

Discussion¶

Related Claims¶

• Apprentice learner models are domain-general: they learn expert models across eight tutor domains without specialized domain knowledge — related
• The review finds empirical evidence that ITS authoring tools benefit non-programmer authors mainly in effectiveness, efficiency, quality of authored artifacts, and usability — related
• Decision trees achieve competitive accuracy with black-box models (LR, RF, SVM), which performed slightly better but lack interpretability — related