Likelihood Guidance
A class of methods for posterior sampling with diffusion models that use the observation likelihood to correct the score during sampling. Such methods typically require likelihood approximation or gradient computation.
Timeline (last 90 days)
The survey categorizes the main methods of diffusion model posterior sampling into four classes: likelihood guidance, spectral and projection methods, sequential Monte Carlo correction, and amortized conditional training.
Likelihood guidance uses the observation likelihood to correct the score during sampling, usually requiring approximation or gradient computation, and may depend on approximation of the likelihood function or gradient availability.
All Facts (2)
The survey categorizes the main methods of diffusion model posterior sampling into four classes: likelihood guidance, spectral and projection methods, sequential Monte Carlo correction, and amortized conditional training.
50%UnverifiedLikelihood guidance uses the observation likelihood to correct the score during sampling, usually requiring approximation or gradient computation, and may depend on approximation of the likelihood function or gradient availability.
50%