Skip to content
BayesLab.Open the app →

BAYESLAB · LEARNING RESOURCES

Teach Bayesian networks with a hands-on browser workspace.

BayesLab is an early-stage teaching tool for educators who want students to build small probabilistic models, test evidence, and explain what changed. Start with one lesson before planning a wider rollout.

By the BayesLab team · Updated 18 September 2026

A suggested 30-minute introductory activity

  1. Predict, 5 minutes. Present the rain and wet-pavement model. Ask students whether wet pavement proves it rained, and why.
  2. Build, 10 minutes. Open BayesLab, add the two variables and their relationship, and enter the probability tables. Check that every row sums to one.
  3. Explore, 10 minutes. Observe Wet, clear the evidence, then observe Dry. Ask students to record their predictions and compare them with the posterior probabilities.
  4. Explain, 5 minutes. Have students explain the role of the prior and alternative explanations. Download the model as a .bayes file to accompany their reasoning.

What students can do today

The editor supports discrete chance nodes, decision and utility nodes, probability-table editing, evidence, worked explanations, and scenario comparison. Built-in examples and a guided introductory tutorial provide starting points. Students can export diagrams and share a snapshot of a model by link.

What a pilot does—and does not—include

Opening the current app requires no installation or account. Submitting the pilot form sends an enquiry to the site administrator; it does not create a student or university account. There are no institutional dashboards, class rosters, LMS connections, or collaborative editing features in this version.

Models are processed in the browser. Autosave belongs to that browser and domain, not a cloud account. Ask students to download a project file when they need a portable copy. A shared-model URL contains the model itself, so anyone receiving that link can read its contents. Use synthetic teaching data and avoid sensitive student or personal information.

Before introducing it to your class

  • Reproduce a representative exercise and independently check the probabilities and expected utility.
  • Try the classroom’s actual devices and browsers, including keyboard navigation and any accessibility tools students use.
  • Check model size and responsiveness; exact inference can become expensive in densely connected networks.
  • Prepare a downloadable starter model and a fallback exercise in case a device or network is unavailable.
  • Review your institution’s requirements before using any tool for assessment or consequential decisions.

Help shape the next version

Tell us your course, the concepts you teach, and what gets in the way of the lesson. We are seeking feedback on usability, explanations, and classroom fit. Register your interest in the university pilot without including student information or sensitive data.

Explore the BayesLab approach to Bayesian modelling, or find your next lesson in our blogs and walkthroughs.