L1-L1 IRLS

[mrava87]

Feature

The IRLS algorithm can handle L1 norm for the data misfit term as well as L1 norm for the model. However, in some scenarios users may want to have both terms together (problem with outliers in the data and sparse model in some domain). The IRLS algorithm can be adapted to handle this scenario by writing the L1-L1 functional as an augmented system of equations and solve it using the classical data-term IRLS.

Implementation

@jdromerom will implement this inpylops.optimization.sparsity.IRLS