Bernstein-von Mises theorems for statistical inverse problems. I: Schrödinger equation (Q783749)

Summary: The inverse problem of determining the unknown potential \(f > 0\) in the partial differential equation \[ \frac{\Delta}{2} u - fu =0 \: \text{ on } \mathcal O, \quad u = g \text{ on } \partial \mathcal O, \] where \(\mathcal O\) is a bounded \(C^\infty\)-domain in \(\mathbb R^d\) and \(g > 0\) is a given source function, is considered. The data consist of the solution \(u\) corrupted by additive Gaussian noise. A nonparametric Bayesian prior for the function \(f\) is devised and a Bernstein-von Mises theorem is proved which entails that the posterior distribution given the observations is approximated by an infinite-dimensional Gaussian measure that has a `minimal' covariance structure in an information-theoretic sense. The function space in which this approximation holds true is shown to carry the finest topology permitted for such a result to be possible. As a consequence the posterior distribution performs valid and optimal frequentist statistical inference on \(f\) in the small noise limit.