Abstract:In this paper we show how to accelerate randomized coordinate descent methods and achieve faster convergence rates without paying per-iteration costs in asymptotic running time. In particular, we show how to generalize and efficiently implement a method proposed by Nesterov, giving faster asymptotic running times for various algorithms that use standard coordinate descent as a black box. In addition to providing a proof of convergence for this new general method, we show that it is numerically stable, efficiently implementable, and in certain regimes, asymptotically optimal. To highlight the power of this algorithm, we show how it can used to create faster linear system solvers in several regimes: • We show how this method achieves a faster asymptotic runtime than conjugate gradient for solving a broad class of symmetric positive definite systems of equations. • We improve the convergence guarantees for Kaczmarz methods, a popular technique for image reconstruction and solving overdetermined systems of equations, by accelerating an algorithm of Strohmer and Vershynin. • We achieve the best known running time for solving Symmetric Diagonally Dominant (SDD) system of equations in the unit-cost RAM model, obtaining a running time of ${\bf O}({\bf m}\log^{\wedge}\{3/2\}{\bf n}(\log\log {\bf n})^{\wedge}\{1/2\}\log((\log {\bf n})/{\bf eps}))$ by accelerating a recent solver by Kelner et al. We show how this method achieves a faster asymptotic runtime than conjugate gradient for solving a broad class of symmetric positive definite systems of equations. We improve the convergence guarantees for Kaczmarz methods, a popular technique for image reconstruction and solving overdetermined systems of equations, by accelerating an algorithm of Strohmer and Vershynin. We achieve the best known running time for solving Symmetric Diagonally Dominant (SDD) system of equations in the unit-cost RAM model, obtaining a running time of ${\bf O}({\bf m}\log^{\wedge}\{3/2\}{\bf n}(\log\log {\bf n})^{\wedge}\{1/2\}\log((\log {\bf n})/{\bf eps}))$ by accelerating a recent solver by Kelner et al. Beyond the independent interest of these solvers, we believe they highlight the versatility of the approach of this paper and we hope that they will open the door for further algorithmic improvements in the future.

Generalized Gearhart-Koshy acceleration for the Kaczmarz method

Acceleration of Randomized Kaczmarz Method via the Johnson-Lindenstrauss Lemma

Accelerated Kaczmarz Algorithms using History Information

An Accelerated Block Randomized Kaczmarz Method

Randomized Kaczmarz method with adaptive stepsizes for inconsistent linear systems

On adapting Nesterov's scheme to accelerate iterative methods for linear problems

On block Gaussian sketching for the Kaczmarz method

Preasymptotic convergence of randomized Kaczmarz method

An accelerated randomized Kaczmarz algorithm

Faster Randomized Block Kaczmarz Algorithms

Randomized Extended Kaczmarz for Solving Least-Squares

Paved with good intentions: Analysis of a randomized block Kaczmarz method

On the adaptive deterministic block Kaczmarz method with momentum for solving large-scale consistent linear systems

Randomized block Kaczmarz method with projection for solving least squares

Efficient Accelerated Coordinate Descent Methods and Faster Algorithms for Solving Linear Systems

Randomized Kaczmarz with geometrically smoothed momentum

Convergence Rates for Greedy Kaczmarz Algorithms, and Faster Randomized Kaczmarz Rules Using the Orthogonality Graph

A Deterministic Kaczmarz Algorithm for Solving Linear Systems

Faster randomized block sparse Kaczmarz by averaging

Randomized Kaczmarz with tail averaging