Machine Learning / Evaluated demonstration
Recommend trials using Bayesian optimisation and resource limits.
Actual recording · Explanatory English captions · Original synthetic data
Recorded demonstration of a functional local application. A pilot defines data, access and deployment; this website serves the video and screenshots.
THE PROBLEM
THE WORKFLOW
A Gaussian process learns from observed trials for a fictional recycled-fibre panel. Explore prediction, uncertainty and expected improvement. Set binder and pressure limits, then run a simulated trial or record your own result. The simulated objective is explicitly identified.
RECORDED EVIDENCE
| Check | Cases |
|---|---|
| Documented checks; synthetic data | 10 of 10 |
Ten heldout seeds, six initial trials and twelve proposals per run. Retrospective checks ask whether the best design is within six score points of the synthetic reference optimum. All ten runs pass; a random search with the same budget is also recorded. Timings cover recommendation, simulated trial and model refitting on CPU, excluding HTTP.
WHAT WE COULD BUILD WITH YOU
Design variables, costs, constraints, measurable objective and how outcomes are recorded.
Synthetic objective, two variables and a discrete candidate grid; this does not demonstrate improvements in real materials. Model uncertainty is not a guarantee. Imported CSV outcomes are used as supplied; simulation is disabled until reset.
Scope a Machine Learning pilot ↗
Explore demo
Sensors → prediction → detected change
Separate changing inputs from declining model performance.
30 of 30 checks · Original synthetic dataDocumented checks; synthetic data
Explore demo
Spectrum → components → unfamiliar signal
Separate spectral components and highlight signals the model cannot explain.
99 of 100 checks · Original synthetic dataDocumented checks; synthetic data