🙌This work is done with the amazing team: Haocheng Ren (co-lead), Gregory Croisdale, Anhong Guo (@anhongguo.bsky.social), and Xu Wang (@xuwanghci.bsky.social).

Rubikon shows the promises of integrating cognitive tutoring principles with mixed reality interfaces. We aim to extend these concepts to broader physical task domains, enabling enhanced personalized learning.
#DIS2025 Paper: arxiv.org/abs/2503.12619
Video: youtu.be/4Jes_0nd3kY

In a study with 36 Rubik’s Cube novices, we found that learners using Rubikon achieved 25% higher learning gains than those using a video tutorial and Rubik’s Cube, as Rubikon increased both the quantity and variety of skill practice while reducing the overhead of preparation.

Rubikon interface: Learners perceive a video stream of themselves manipulating the cube, a goal image for the current objective, a skillometer of knowledge mastery, and adaptive functionalities like extended views for hidden faces and hints at different levels of detail.

Rubikon’s core idea is thus to enable quick, targeted practice based on learner weaknesses. Building on cognitive tutor principles and instantiating an AR setup, Rubikon tracks learner progress, assesses skill mastery, and generates new cube configurations for repeated practice.

Why is learning to solve a Rubik’s Cube challenging? Its solution involves multiple stages each with distinct knowledge components. With passive media, novice learners often miss certain cases even after multiple passes yet cannot easily reconfigure the cube to practice them.

Rubikon shows the promises of integrating cognitive tutoring principles with mixed reality interfaces. We aim to extend these concepts to broader physical task domains, enabling enhanced personalized learning.
#DIS2025 Paper: arxiv.org/abs/2503.12619
Video: youtu.be/4Jes_0nd3kY

In a study with 36 Rubik’s Cube novices, we found that learners using Rubikon achieved 25% higher learning gains than those using a video tutorial and Rubik’s Cube, as Rubikon increased both the quantity and variety of skill practice while reducing the overhead of preparation.

Rubikon interface: Learners perceive a video stream of themselves manipulating the cube, a goal image for the current objective, a skillometer of knowledge mastery, and adaptive functionalities like extended views for hidden faces and hints at different levels of detail.

Rubikon’s core idea is thus to enable quick, targeted practice based on learner weaknesses. Building on cognitive tutor principles and instantiating an AR setup, Rubikon tracks learner progress, assesses skill mastery, and generates new cube configurations for repeated practice.

Why is learning to solve a Rubik’s Cube challenging? Its solution involves multiple stages each with distinct knowledge components. With passive media, novice learners often miss certain cases even after multiple passes yet cannot easily reconfigure the cube to practice them.