Uncertainty Quantification with Generative Models
This work addresses uncertainty quantification in image reconstruction for researchers and practitioners, offering a method that is efficient and broadly applicable but incremental in its approach.
The paper tackles Bayesian inverse problems like image reconstruction by using generative models to incorporate complex data-driven priors and enable efficient uncertainty quantification, achieving computational tractability and applicability to arbitrary corruption types after a single training phase.
We develop a generative model-based approach to Bayesian inverse problems, such as image reconstruction from noisy and incomplete images. Our framework addresses two common challenges of Bayesian reconstructions: 1) It makes use of complex, data-driven priors that comprise all available information about the uncorrupted data distribution. 2) It enables computationally tractable uncertainty quantification in the form of posterior analysis in latent and data space. The method is very efficient in that the generative model only has to be trained once on an uncorrupted data set, after that, the procedure can be used for arbitrary corruption types.