On statistical Calderón problems

Abstract: For

D

a bounded domain in

m a t h b b R^{d}, d g e 2,

with smooth boundary

p a r t i a l D

, the non-linear inverse problem of recovering the unknown conductivity

g a m m a

determining solutions

u = u_{g a m m a, f}

of the partial differential equation �egin{equation*} �egin{split}

abla cdot(gamma abla u)&=0 quad ext{ in }D, \ u&=f quad ext { on } partial D, end{split} end{equation*} from noisy observations $Y$ of the Dirichlet-to-Neumann map [f mapsto Lambda_gamma(f) = {gamma frac{partial u_{gamma,f}}{partial

u}}Big|_{partial D},] with

p a r t i a l / p a r t i a l u

denoting the outward normal derivative, is considered. The data

Y

consists of

L a m b d a_{g} a m m a

corrupted by additive Gaussian noise at noise level

v a r e p s i l o n > 0

, and a statistical algorithm

h a t g a m m a (Y)

is constructed which is shown to recover

g a m m a

in supremum-norm loss at a statistical convergence rate of the order

l o g (1 / v a r e p s i l o n)^{- d e l t a}

as

v a r e p s i l o n o 0

. It is further shown that this convergence rate is optimal, up to the precise value of the exponent

d e l t a > 0

, in an information theoretic sense. The estimator

h a t g a m m a (Y)

has a Bayesian interpretation in terms of the posterior mean of a suitable Gaussian process prior and can be computed by MCMC methods.

Review text

Summary: For $D$ a bounded domain in $\mathbb{R}^d$, $d \geq 2$, with smooth boundary $\partial D$, the non-linear inverse problem of recovering the unknown conductivity $\gamma$ determining solutions $u=u_{\gamma, f}$ of the partial differential equation \[ \begin{aligned} \nabla \cdot(\gamma \nabla u) &=0 \quad \text{in } D, \\ u&=f \quad \text{on } \partial D, \end{aligned} \] from noisy observations $Y$ of the Dirichlet-to-Neumann map \[f \mapsto \Lambda_\gamma(f) = \gamma \frac{\partial u_{\gamma,f}}{\partial \nu}\Big|_{\partial D},\] with $\partial/\partial \nu$ denoting the outward normal derivative, is considered. The data $Y$ consists of $\Lambda_\gamma$ corrupted by additive Gaussian noise at noise level $\epsilon>0$, and a statistical algorithm $\hat{\gamma}(Y)$ is constructed which is shown to recover $\gamma$ in supremum-norm loss at a statistical convergence rate of the order $\log(1/\epsilon)^{-\delta}$ as $\epsilon \to 0$. It is further shown that this convergence rate is optimal, up to the precise value of the exponent $\delta>0$, in an information theoretic sense. The estimator $\hat{\gamma}(Y)$ has a Bayesian interpretation in terms of the posterior mean of a suitable Gaussian process prior and can be computed by MCMC methods.

Recommendations

Cites work

Cited in

(43)

This page was built for publication: On statistical Calderón problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q778889)