An Efficient Semismooth Newton Based Algorithm for Convex Clustering

被引:0
作者
Yuan, Yancheng [1 ]
Sun, Defeng [2 ]
Toh, Kim-Chuan [1 ]
机构
[1] Natl Univ Singapore, Dept Math, Singapore, Singapore
[2] Hong Kong Polytech Univ, Dept Appl Math, Hong Kong, Peoples R China
来源
INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 80 | 2018年 / 80卷
关键词
AUGMENTED LAGRANGIAN METHOD;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Clustering is a fundamental problem in unsupervised learning. Popular methods like K-means, may suffer from instability as they are prone to get stuck in its local minima. Recently, the sumof-norms (SON) model (also known as clustering path), which is a convex relaxation of hierarchical clustering model, has been proposed in (Lindsten et al., 2011) and (Hocking et al., 2011). Although numerical algorithms like alternating direction method of multipliers (ADMM) and alternating minimization algorithm (AMA) have been proposed to solve convex clustering model (Chi & Lange, 2015), it is known to be very challenging to solve large-scale problems. In this paper, we propose a semismooth Newton based augmented Lagrangian method for large-scale convex clustering problems. Extensive numerical experiments on both simulated and real data demonstrate that our algorithm is highly efficient and robust for solving large-scale problems. Moreover, the numerical results also show the superior performance and scalability of our algorithm comparing to existing first-order methods.
引用
收藏
页数:9
相关论文
共 24 条
[1]  
[Anonymous], 2016, ARXIV160206612
[2]   Relax, No Need to Round: Integrality of Clustering Formulations [J].
Awasthi, Pranjal ;
Bandeira, Afonso S. ;
Charikar, Moses ;
Krishnaswamy, Ravishankar ;
Villar, Soledad ;
Ward, Rachel .
PROCEEDINGS OF THE 6TH INNOVATIONS IN THEORETICAL COMPUTER SCIENCE (ITCS'15), 2015, :191-200
[3]   An efficient inexact symmetric Gauss-Seidel based majorized ADMM for high-dimensional convex composite conic programming [J].
Chen, Liang ;
Sun, Defeng ;
Toh, Kim-Chuan .
MATHEMATICAL PROGRAMMING, 2017, 161 (1-2) :237-270
[4]   Splitting Methods for Convex Clustering [J].
Chi, Eric C. ;
Lange, Kenneth .
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS, 2015, 24 (04) :994-1013
[5]  
Clarke F., 1983, OPTIMIZATION NONSMOO
[6]   From few to many: Illumination cone models for face recognition under variable lighting and pose [J].
Georghiades, AS ;
Belhumeur, PN ;
Kriegman, DJ .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2001, 23 (06) :643-660
[7]  
GEORGHIADES AS, 2001, YALE FACE DATABASE
[8]   GENERALIZED HESSIAN MATRIX AND 2ND-ORDER OPTIMALITY CONDITIONS FOR PROBLEMS WITH C1,1 DATA [J].
HIRIARTURRUTY, JB ;
STRODIOT, JJ ;
NGUYEN, VH .
APPLIED MATHEMATICS AND OPTIMIZATION, 1984, 11 (01) :43-56
[9]  
Hocking T.D., 2011, PROC 28 INT C MACHIN, P745
[10]  
Lecun Yann, 1998, Mnist dataset