Skip to main content
Esperanto rainbow question from quiz

BKT Explainable with Hypothesis Generation

Summer 2019. The post-hoc explainable for Bayesian Knowledge Tracing was accepted to the 2019 IEEE Workshop on Visualization for AI Explainability. [Noah Cowit '20]

a hot air balloon with parameter sliders

BKT Explainable with Hypothesis Generation

Summer 2019. Using a hot air balloon metaphor, you can interact with the parameters of Bayesian Knowledge Tracing and see how it influences the model's predictions! [Catherine Yeh '22]

Paper prototyping process image

What is Bayesian Knowledge Tracing?

This overview of our Summer 2018 BKT Explainables was accepted as a poster to the 2018 IEEE Workshop on Visualization for AI Explainability.

Screenshot of the alchemy BKT explainable

Static vs. Interactive Alchemy BKT Explainable

2018-2020. In our paper in AAAI AIES 2020 "Assessing Post-hoc Explainability of the BKT Algorithm", we compare the difference in learning gains of Mechanical Turk workers using a static or interactive version of the Alchemy BKT Explainable from Summer 2018. [Haoyu Sheng '20, Tongyu Zhou '20]

A screenshot of a ghost learning to cook

Cooking BKT Explainable

Summer 2018. This project uses a story about learning to cook to explain the various parts of the BKT algorithm. [Grace Mazzarella '19]

A driver deciding between two options

Paper Prototype of Road Trip BKT Explainable

Summer 2018. This is a paper prototype of a driving analogy to explain the various components of the Bayesian Knowledge Tracing Algorithm. [Young Cho '19]

Screenshot of the alchemy BKT explainable

Alchemy BKT Explainable

Summer 2018. This is an early prototype of the Alchemy BKT Explainable featured in our paper in AAAI AIES 2020 "Assessing Post-hoc Explainability of the BKT Algorithm". Later, a static and interactive version was created and evaluated on Mechanical Turk. [Kelvin Tejeda '20, Tongyu Zhou '20]

Screenshot of the EEP BKT system

Experimental Educational Platform (EEP)

Summer 2018. The EEP is a Bayesian Knowledge Tracing System to provide context in experiments for participants of our BKT Explainability studies. It is written in Javascript and uses a node.js server for persistent data. [Kelvin Tejeda '20]