We review a variety of iterative methods for inference and estimation in Gaussian graphical models, also known as Gauss-Markov random fields (GMRFs), and consider how to adapt these methods to non-lin
