A Major Update for HUGE and SAM

With the recent progress in the theoretical field of sparse learning problems, current R packages are lagging behind the cutting edge research. We aim to update the current R package in order to achieve the state-of-the-art performance and equip them with various functionalities. Specifically, we will update the algorithms for estimating (1) sparse undirected graphical model with a novel active-set based second-order optimization scheme, and (2) nonparametric regression and classification model with a Newton-type blockwise coordinate descent algorithm. Furthermore, we will add inference module and sparsity induced regularization functions to the packages.