Diffusion & denoising in 3D
Compare forward noise with learned reverse denoising. Explore the 3D model, inspect its parts and learn with guided explanations. Free on Atlas.
What you can learn
Compare forward noise with learned reverse denoising.
Start with a structured sample.
Corrupt in controlled steps.
Condition a learned prediction on the step.
Try an investigation
Trace clean example toward forward noising equation. Which intermediate operations or assemblies are essential, and which assumptions are simplified by this diagram?
Check your understanding
Forward diffusion adds…
Controlled noise. Controlled noise. The exhibit shows computational roles and dependencies; it does not contain trained parameters or run inference.
The denoiser uses the noisy sample and…
Noise-step information. Noise-step information. The exhibit shows computational roles and dependencies; it does not contain trained parameters or run inference.
Reverse generation samples…
A learned distribution. A learned distribution. The exhibit shows computational roles and dependencies; it does not contain trained parameters or run inference.
Sources and model scope
A three-dimensional computational diagram, not physical hardware or a running trained model. Displayed widths, node counts and example values are illustrative. Select named stages to trace information flow. The separate upper study rail contains deterministic worked examples, with values on every selectable cell.