Guided diffusion methods reveal actual data resolution in reconstructions
Measuring Effective Data Resolution in Guided Diffusion Posteriors
Machine Learning
Summary
Sometimes when we use computers to fill in missing information from limited measurements, it’s hard to tell how much the computer's guess really depends on the data versus its own assumptions. The authors offer a way to measure how finely the data guides these guesses, called effective data resolution. They test this measurement method and use it to show that certain settings in guided diffusion can change how much data influences the result, in ways not obvious from usual checks. Their work helps clarify when and how guided diffusion samplers truly reflect the original data.
What this means in practice
- •For environmental monitoring teams: Assess how much sparse sensor data actually influences high-resolution field reconstructions using guided diffusion methods.
- •For medical imaging engineers: Quantify the true resolution contribution of limited scan data in diffusion-based image reconstruction workflows.
Authors
Ridham Patel, Defu Cao, Jiacheng Pang, Yan Liu
Abstract
Guided diffusion samplers are increasingly used to reconstruct physical fields from sparse observations, but standard diagnostics do not say how much of the reconstruction was actually determined by the data. We introduce effective data resolution for black-box generative posteriors: a comparison between the resolution warranted by the inverse problem, $\mathrm{dof}_{\mathrm{ref}}$, and the resolution realised by the sampler, $\mathrm{dof}_{\mathrm{samp}}$. A perturbation estimator measures $\mathrm{dof}_{\mathrm{samp}}$ and the spatial map $R(x,x)$ from sampler queries alone. We validate the estimator against exact references and use it to study guided diffusion. The resulting measurements show that guidance weight can strongly alter apparent information transfer, that mean, spread and resolution are not jointly corrected by one weight even with an exact prior and score, and that resolution fidelity does not follow reliably from the apparent principledness of a guidance rule.