Skip to main content

Explainable by design

A model that is its own explanation.

Upload your data. Get a model you can read.

Diafunc discovers a short, explicit model from your table with Symbolic Regression, so there is nothing to explain after the fact.

Explanation tools describe a trained model from the outside, with charts about what mattered. Here the explanation and the model are the same object: when someone asks why, you show them the model, with the evidence behind every number.

What you get
  • · A model you can read
  • · Its quality, measured on rows it never saw
  • · A full report in plain language
  • · Your questions, answered
  • · Ready to use: as code, as an API, as a PDF report

Runs from about an hour. We email you when it is done.

The chapter "How each term responds to its inputs" of the Yacht Hydrodynamics report: the formula is a constant and one term that reads all three of its inputs, and that term is traced over Froude number, over Longitudinal position of the center of buoyancy and over Prismatic coefficient, one curve per held-out row; the curves over Froude number rise ever more steeply, the curves over the center of buoyancy rise slightly, the curves over Prismatic coefficient fall slightly.
Real output. The model traced over each of its inputs: the explanation is the model itself.

50,000 free credits to start: about twelve analyses at Standard depth. No credit card.

Hosted in the EU Private by default Cancel anytime

Start free