Parallel Coordinate Descent Newton Method for Efficient $\ell_1$-Regularized Minimization

7 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

CoCoA: A General Framework for Communication-Efficient Distributed Optimization (2016, methodology): “, solvers based on the primal formulation of L1-regularized objectives (Shalev-Shwartz and Tewari, 2011; Yuan et al., 2012; Bian et al., 2013; Fercoq and Richtárik, 2015; Tappenden et al., 2015).”
A distributed block coordinate descent method for training $l_1$ regularized linear classifiers (2014, methodology): “Recall from subsection 4.5 that PCD-S is the variation of the PCDN method Bian et al. (2013) using the S-scheme for variable selection.”
Hessian-CoCoA : a general parallel and distributed framework for non-strongly convex regularizers (2017, background): “Despite that, what the literature currently offers is not satisfying yet the requirements: focusing exclusively on parallel methods over features, the state-of-the-art is represented by Shotgun Coordinate Descent Newton, described in [5], [15], that originates from the ineherently sequential stochas”
L1-Regularized Distributed Optimization: A Communication-Efficient Primal-Dual Framework (2015, methodology): “Parallel coordinate descent for L1-regularized objectives (with and without using mini-batches) was proposed in [7] (Shotgun) and generalized in [3] , and is among the best performing solvers in the parallel setting.”