Scaling Up Coordinate Descent Algorithms for Large $\ell_1$ Regularization Problems

Ambuj Tewari (University of Texas); Chad Scherrer (Pacific Northwest National Lab); David Haglin (Pacific Northwest National Lab); Mahantesh Halappanavar (Pacific Northwest National Lab)

arxiv: 1206.6409 · v1 · pith:DJQS4POJnew · submitted 2012-06-27 · 💻 cs.LG · cs.DC· stat.ML

Scaling Up Coordinate Descent Algorithms for Large ell₁ Regularization Problems

Chad Scherrer (Pacific Northwest National Lab) , Mahantesh Halappanavar (Pacific Northwest National Lab) , Ambuj Tewari (University of Texas) , David Haglin (Pacific Northwest National Lab) This is my paper

classification 💻 cs.LG cs.DCstat.ML

keywords algorithmsparallelcoordinatedescentspecialalgorithmcasescases---thread-greedy

0 comments

read the original abstract

We present a generic framework for parallel coordinate descent (CD) algorithms that includes, as special cases, the original sequential algorithms Cyclic CD and Stochastic CD, as well as the recent parallel Shotgun algorithm. We introduce two novel parallel algorithms that are also special cases---Thread-Greedy CD and Coloring-Based CD---and give performance measurements for an OpenMP implementation of these.

This paper has not been read by Pith yet.

Scaling Up Coordinate Descent Algorithms for Large ell₁ Regularization Problems

discussion (0)