Convergent regularization of an inverse problem

A regularization $R_{\alpha }$ is convergent when its parameter can be chosen from the noise level and data so that the regularized reconstruction converges to the exact Moore–Penrose inverse of an operator solution as the noise level tends to zero.