图 书 馆
关于凸凹过程的收敛性On the Convergence of the Convex-Concave Procedure |
|
| 课程网址: | http://videolectures.net/nipsworkshops09_sriperumbudur_ccc/ |
| 主讲教师: | Bharath K. Sriperumbudur |
| 开课单位: | 圣地亚哥大学 |
| 开课时间: | 2010-01-19 |
| 课程语种: | 英语 |
| 中文简介: |  |
| 课程简介: | The concave-convex procedure (CCCP) is a majorization-minimization algorithm that solves d.c. (difference of convex functions) programs as a sequence of convex programs. In machine learning, CCCP is extensively used in many learning algorithms like sparse support vector machines (SVMs), transductive SVMs, sparse principal component analysis, etc. Though widely used in many applications, the convergence behavior of CCCP has not gotten a lot of specific attention. In this paper, we provide a rigorous analysis of the convergence of CCCP by addressing these questions: *(i) When does CCCP find a local minimum or a stationary point of the d.c. program under consideration? *(ii) When does the sequence generated by CCCP converge? We also present an open problem on the issue of local convergence of CCCP. |
| 关 键 词: | 计算机科学; 机器学习; 核方法; 支持向量机 |
| 课程来源: | 视频讲座网 |
| 最后编审: | 2020-06-11：liush |
| 阅读次数: | 111 |