Reproduce “Frac-BKT Under Mechanism Scrutiny: Response Assimilation and a Failed Fractional-Memory Test” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Nizam Kadir

Frac-BKT predicts responses from prior mastery states, then updates each state only after scoring the observed response. Its fractional-history term adds weighted prior state changes through an elapsed-time-scaled approximation, not an exact irregular-time solver. On ASSISTments 2017, the repaired model reaches .5860 overall AUC, but removing history improves AUC by about .009 on both delayed next-response slices; these are not direct retention tests.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.