Diffusion Posterior Sampling
A class of methods that use a pretrained diffusion model as a prior and, at inference time, combine it with a measurement operator for Bayesian posterior sampling. Reconstruction is guided by decomposing the prior score and the likelihood score.
Timeline (last 90 days)
Diffusion posterior sampling is a class of methods that uses a pretrained diffusion model as a prior and incorporates the measurement operator at inference time.
Existing diffusion posterior sampling methods mostly use approximate computation of the likelihood score, for example substituting the gradient of the measurement residual.
All Facts (2)
Diffusion posterior sampling is a class of methods that uses a pretrained diffusion model as a prior and incorporates the measurement operator at inference time.
50%UnverifiedExisting diffusion posterior sampling methods mostly use approximate computation of the likelihood score, for example substituting the gradient of the measurement residual.
50%