期刊
MATHEMATICAL PROGRAMMING
卷 151, 期 1, 页码 3-34出版社
SPRINGER HEIDELBERG
DOI: 10.1007/s10107-015-0892-3
关键词
Coordinate descent; Randomized algorithms; Parallel numerical computing
类别
资金
- NSF [DMS-1216318, IIS-1447449]
- ONR [N00014-13-1-0129]
- AFOSR from Argonne National Laboratory [FA9550-13-1-0138, 3F-30222]
- Direct For Mathematical & Physical Scien
- Division Of Mathematical Sciences [1216318] Funding Source: National Science Foundation
- Div Of Information & Intelligent Systems
- Direct For Computer & Info Scie & Enginr [1447449] Funding Source: National Science Foundation
Coordinate descent algorithms solve optimization problems by successively performing approximate minimization along coordinate directions or coordinate hyperplanes. They have been used in applications for many years, and their popularity continues to grow because of their usefulness in data analysis, machine learning, and other areas of current interest. This paper describes the fundamentals of the coordinate descent approach, together with variants and extensions and their convergence properties, mostly with reference to convex objectives. We pay particular attention to a certain problem structure that arises frequently in machine learning applications, showing that efficient implementations of accelerated coordinate descent algorithms are possible for problems of this type. We also present some parallel variants and discuss their convergence properties under several models of parallel execution.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据