In this paper, we consider solving the robust linear regression problem y = Ax + ∈ by an inexact Newton method and an iteratively reweighted least squares method. We show that each of these methods can be combined with ...