Ensemble Kalman inversion (EKI) is a recent technique for the treatment of high-dimensional inverse problems without evaluating the adjoint of the forward operator.
In this talk, we will look at EKI for linear inverse problems from the perspective of classical, functional-analytic regularization theory. In particular, convergence rates under standard source conditions can be shown if the sample size is iteratively adapted to the regularization parameter. Further practical aspects are also discussed.