Well-posed Bayesian Inverse Problems: Priors with Exponential Tails
This provides theoretical foundations for Bayesian inverse problems with non-Gaussian priors, benefiting practitioners in inverse problems and uncertainty quantification.
The paper establishes well-posedness (existence, uniqueness, stability) for Bayesian inverse problems with convex priors having exponential tails, including Gaussian and Besov measures, under specific conditions on the likelihood. It also provides a recipe for constructing such priors on Banach spaces.
We consider the well-posedness of Bayesian inverse problems when the prior measure has exponential tails. In particular, we consider the class of convex (log-concave) probability measures which include the Gaussian and Besov measures as well as certain classes of hierarchical priors. We identify appropriate conditions on the likelihood distribution and the prior measure which guarantee existence, uniqueness and stability of the posterior measure with respect to perturbations of the data. We also consider consistent approximations of the posterior such as discretization by projection. Finally, we present a general recipe for construction of convex priors on Banach spaces which will be of interest in practical applications where one often works with spaces such as $L^2$ or the continuous functions.