Reproduce “Model Human Learners: Computational Models to Guide Instructional Design” 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

Christopher J. MacLellan

This work demonstrates a computational learning model's ability to accurately predict outcomes from human A/B experiments in instructional design. By simulating diverse instructional interventions *in silico*, the Model Human Learner concept provides a more efficient and theoretically informed approach to optimizing educational materials.

Key insights

  • Computational models can reliably predict main effects observed in human A/B experiments, demonstrating their utility as a design evaluation tool.
  • The model is capable of generating realistic learning curve predictions without prior training on human data, crucial for evaluating novel instructional designs.
  • These models can offer deeper theoretical insights, such as identifying procedural ambiguity as a key factor in instructional effectiveness, challenging existing hypotheses.

Summary by alphaXiv.

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