Autoencoder Constrained Clustering With Adaptive Neighbors

被引:48
作者
Li, Xuelong [1 ,2 ]
Zhang, Rui [1 ,2 ]
Wang, Qi [1 ,2 ]
Zhang, Hongyuan [1 ,2 ]
机构
[1] Northwestern Polytech Univ, Sch Comp Sci, Xian 710072, Peoples R China
[2] Northwestern Polytech Univ, Ctr Opt IMagery Anal & Learning OPTIMAL, Xian 710072, Peoples R China
基金
中国博士后科学基金; 中国国家自然科学基金;
关键词
Kernel; Adaptive systems; Clustering methods; Clustering algorithms; Sparse matrices; Learning systems; Neural networks; Adaptive neighbors; autoencoder; deep clustering; parameter-free similarity; structured graph;
D O I
10.1109/TNNLS.2020.2978389
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The conventional subspace clustering method obtains explicit data representation that captures the global structure of data and clusters via the associated subspace. However, due to the limitation of intrinsic linearity and fixed structure, the advantages of prior structure are limited. To address this problem, in this brief, we embed the structured graph learning with adaptive neighbors into the deep autoencoder networks such that an adaptive deep clustering approach, namely, autoencoder constrained clustering with adaptive neighbors (ACC_AN), is developed. The proposed method not only can adaptively investigate the nonlinear structure of data via a parameter-free graph built upon deep features but also can iteratively strengthen the correlations among the deep representations in the learning process. In addition, the local structure of raw data is preserved by minimizing the reconstruction error. Compared to the state-of-the-art works, ACC_AN is the first deep clustering method embedded with the adaptive structured graph learning to update the latent representation of data and structured deep graph simultaneously.
引用
收藏
页码:443 / 449
页数:7
相关论文
共 34 条
[1]  
[Anonymous], 2016, ARXIV161004794
[2]  
[Anonymous], 2019, IEEE T NEUR NET LEAR
[3]  
Chang JL, 2017, IEEE I CONF COMP VIS, P5880, DOI [10.1109/ICCV.2017.627, 10.1109/ICCV.2017.626]
[4]   Subspace clustering using a low-rank constrained autoencoder [J].
Chen, Yuanyuan ;
Zhang, Lei ;
Yi, Zhang .
INFORMATION SCIENCES, 2018, 424 :27-38
[5]  
Dalal N., 2005, IEEE C COMPUT VIS PA, V1, P886
[6]   A min-max cut algorithm for graph partitioning and data clustering [J].
Ding, CHQ ;
He, XF ;
Zha, HY ;
Gu, M ;
Simon, HD .
2001 IEEE INTERNATIONAL CONFERENCE ON DATA MINING, PROCEEDINGS, 2001, :107-114
[7]   Deep Clustering via Joint Convolutional Autoencoder Embedding and Relative Entropy Minimization [J].
Dizaji, Kamran Ghasedi ;
Herandi, Amirhossein ;
Deng, Cheng ;
Cai, Weidong ;
Huang, Heng .
2017 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), 2017, :5747-5756
[8]   Sparse Subspace Clustering: Algorithm, Theory, and Applications [J].
Elhamifar, Ehsan ;
Vidal, Rene .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2013, 35 (11) :2765-2781
[9]  
Favaro P, 2011, PROC CVPR IEEE, P1801, DOI 10.1109/CVPR.2011.5995365
[10]  
Feng J., P 25 INT JOINT C ART, V2016, P1925