Entropy-Rate Clustering: Cluster Analysis via Maximizing a Submodular Function Subject to a Matroid Constraint

被引:87
|
作者
Liu, Ming-Yu [1 ]
Tuzel, Oncel [1 ]
Ramalingam, Srikumar [1 ]
Chellappa, Rama [2 ]
机构
[1] MERL, Cambridge, MA 02139 USA
[2] Univ Maryland, College Pk, MD 20742 USA
关键词
Clustering; superpixel segmentation; graph theory; information theory; submodular function; discrete optimization; IMAGE; APPROXIMATIONS; SHAPE;
D O I
10.1109/TPAMI.2013.107
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We propose a new objective function for clustering. This objective function consists of two components: the entropy rate of a random walk on a graph and a balancing term. The entropy rate favors formation of compact and homogeneous clusters, while the balancing function encourages clusters with similar sizes and penalizes larger clusters that aggressively group samples. We present a novel graph construction for the graph associated with the data and show that this construction induces a matroid-a combinatorial structure that generalizes the concept of linear independence in vector spaces. The clustering result is given by the graph topology that maximizes the objective function under the matroid constraint. By exploiting the submodular and monotonic properties of the objective function, we develop an efficient greedy algorithm. Furthermore, we prove an approximation bound of 1/2 for the optimality of the greedy solution. We validate the proposed algorithm on various benchmarks and show its competitive performances with respect to popular clustering algorithms. We further apply it for the task of superpixel segmentation. Experiments on the Berkeley segmentation data set reveal its superior performances over the state-of-the-art superpixel segmentation algorithms in all the standard evaluation metrics.
引用
收藏
页码:99 / 112
页数:14
相关论文
共 8 条
  • [1] MAXIMIZING A MONOTONE SUBMODULAR FUNCTION SUBJECT TO A MATROID CONSTRAINT
    Calinescu, Gruia
    Chekuri, Chandra
    Pal, Martin
    Vondrak, Jan
    [J]. SIAM JOURNAL ON COMPUTING, 2011, 40 (06) : 1740 - 1766
  • [2] On maximizing a monotone k-submodular function subject to a matroid constraint
    Sakaue, Shinsaku
    [J]. DISCRETE OPTIMIZATION, 2017, 23 : 105 - 113
  • [3] Maximizing a submodular set function subject to a matroid constraint (Extended abstract)
    Calinescu, Gruia
    Chekuri, Chandra
    Pal, Martin
    Vondrak, Jan
    [J]. INTEGER PROGRAMMING AND COMBINATORIAL OPTIMIZATION, PROCEEDINGS, 2007, 4513 : 182 - +
  • [4] On Maximizing the Difference Between an Approximately Submodular Function and a Linear Function Subject to a Matroid Constraint
    Wang, Yijing
    Xu, Yicheng
    Yang, Xiaoguang
    [J]. COMBINATORIAL OPTIMIZATION AND APPLICATIONS, COCOA 2021, 2021, 13135 : 75 - 85
  • [5] A note on maximizing a submodular set function subject to a knapsack constraint
    Sviridenko, M
    [J]. OPERATIONS RESEARCH LETTERS, 2004, 32 (01) : 41 - 43
  • [6] Maximizing Approximately Non-k-Submodular Monotone Set Function with Matroid Constraint
    Jiang, Yanjun
    Wang, Yijing
    Yang, Ruiqi
    Ye, Weina
    [J]. THEORY AND APPLICATIONS OF MODELS OF COMPUTATION, TAMC 2022, 2022, 13571 : 11 - 20
  • [7] Faster approximation algorithms for maximizing a monotone submodular function subject to a b-matching constraint
    Fujii, Kaito
    [J]. INFORMATION PROCESSING LETTERS, 2016, 116 (09) : 578 - 584
  • [8] Maximizing the Sum of a Supermodular Function and a Monotone DR-submodular Function Subject to a Knapsack Constraint on the Integer Lattice
    Tan, Jingjing
    Xu, Yicheng
    Zhang, Dongmei
    Zhang, Xiaoqing
    [J]. COMPUTATIONAL DATA AND SOCIAL NETWORKS, CSONET 2021, 2021, 13116 : 68 - 75