Distributed Coordinate Descent Method for Learning with Big Data
Peter Richtárik, Martin Takáč · Edinburgh Research Explorer (University of Edinburgh) · 2013
In this paper we develop and analyze Hydra: HYbriD cooRdinAte descent method for solving loss minimization problems with big data. We initially partition the coordinates (features) and assign each partition to a different node of a cluster. At every iteration, each node picks a random subset of the coordinates from those it owns, independently from the other computers, and in parallel computes and applies updates to the selected coordinates based on a simple closed-form formula. We give bounds on the number of iterations sufficient to approximately solve the problem with high probability, and show how it depends on the data and on the partitioning. We perform numerical experiments with a LASSO instance described by a 3TB matrix.
Cited by 182 works. doi:10.48550/arxiv.1310.2059
Authors
References
- Pegasos: primal estimated sub-gradient solver for SVM
- A dual coordinate descent method for large-scale linear SVM
- Gradient methods for minimizing composite objective function
- Efficiency of Coordinate Descent Methods on Huge-Scale Optimization Problems
- Stochastic dual coordinate ascent methods for regularized loss
- Iteration complexity of randomized block-coordinate descent methods for minimizing a composite function
- Stochastic methods for l 1 regularized loss minimization
- Parallel coordinate descent methods for big data optimization
- Mini-Batch Primal and Dual Methods for SVMs
- Parallel Coordinate Descent for L1-Regularized Loss Minimization