Proposal: Graphical lasso #59

I'd like to formalize theorems around graphical lasso which corresponds to Section 11.2 in high dimensional statistics.
Instead of working through a textbook, I would like to formalize the results of "High-dimensional covariance estimation by minimizing ℓ1-penalized log-determinant divergence" step by step.