Adaptive Graph K-Means

被引:0
作者
Pei, Shenfei [1 ]
Sun, Yuanchen [2 ]
Nie, Feiping [3 ]
Jiang, Xudong [4 ]
Zheng, Zengwei [1 ]
机构
[1] Hangzhou City Univ, Sch Comp & Comp Sci, Hangzhou 310015, Zhejiang, Peoples R China
[2] Zhejiang Univ, Coll Comp Sci & Technol, Hangzhou 310027, Zhejiang, Peoples R China
[3] Northwestern Polytech Univ, Sch Comp Sci, Xian 710072, Shaanxi, Peoples R China
[4] Nanyang Technol Univ, Sch Elect & Elect Engn, Singapore 639798, Singapore
基金
中国国家自然科学基金;
关键词
Machine learning; Clustering; Graph-based; k-means; Computational efficiency;
D O I
10.1016/j.patcog.2024.111226
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Clustering large-scale datasets has received increasing attention recently. However, existing algorithms are still not efficient in scenarios with extremely large number of clusters. To this end, Adaptive Graph K-Means (AGKM) is proposed in this work. Its idea originates from k-means, but it operates on an adaptive k-Nearest Neighbor (k-NN) graph instead of data features. First, AGKM is highly efficient for processing datasets where both the numbers of samples and clusters are very large. Specifically, the time and space complexity are both linear w.r.t the number of samples and, more importantly, independent to the cluster number. Second, AGKM is designed for balanced clusters. This constraint is realized by adding a regularization term in loss function, and a simple modification of the graph in optimization algorithm, which does not increase the computational burden. Last, the indicator and dissimilarity matrices are learned simultaneously, so that the proposed AGKM obtains the final partition directly with higher efficacy and efficiency. Experiments on several datasets validate the advantages of AGKM. In particular, over 29X and 46X speed-ups with respect to k-means are observed on the two large-scale datasets WebFace and CelebA, respectively.
引用
收藏
页数:11
相关论文
共 50 条
[41]   t-k-means: A ROBUST AND STABLE k-means VARIANT [J].
Li, Yiming ;
Zhang, Yang ;
Tang, Qingtao ;
Huang, Weipeng ;
Jiang, Yong ;
Xia, Shu-Tao .
2021 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP 2021), 2021, :3120-3124
[42]   QuicK-means: accelerating inference for K-means by learning fast transforms [J].
Giffon, Luc ;
Emiya, Valentin ;
Kadri, Hachem ;
Ralaivola, Liva .
MACHINE LEARNING, 2021, 110 (05) :881-905
[43]   QuicK-means: accelerating inference for K-means by learning fast transforms [J].
Luc Giffon ;
Valentin Emiya ;
Hachem Kadri ;
Liva Ralaivola .
Machine Learning, 2021, 110 :881-905
[44]   K-Means Clustering Based on Self-adaptive Weight [J].
Zhang, Yuzhu ;
Shi, Hualin ;
Zhang, Damin .
PROCEEDINGS OF 2012 2ND INTERNATIONAL CONFERENCE ON COMPUTER SCIENCE AND NETWORK TECHNOLOGY (ICCSNT 2012), 2012, :1540-1544
[45]   Single Image Super Resolution by Adaptive K-means Clustering [J].
Rahnama, Javad ;
Shirpour, Mohsen ;
Manzuri, Mohammad Taghi .
2017 10TH IRANIAN CONFERENCE ON MACHINE VISION AND IMAGE PROCESSING (MVIP), 2017, :209-214
[46]   Online K-Means Clustering With Adaptive Dual Cost Functions [J].
Sharma, Priti ;
Sharma, Amit .
2017 INTERNATIONAL CONFERENCE ON INTELLIGENT COMPUTING, INSTRUMENTATION AND CONTROL TECHNOLOGIES (ICICICT), 2017, :793-799
[47]   Clustering of Image Data Using K-Means and Fuzzy K-Means [J].
Rahmani, Md. Khalid Imam ;
Pal, Naina ;
Arora, Kamiya .
INTERNATIONAL JOURNAL OF ADVANCED COMPUTER SCIENCE AND APPLICATIONS, 2014, 5 (07) :160-163
[48]   Improving Clustering Method Performance Using K-Means, Mini Batch K-Means, BIRCH and Spectral [J].
Wahyuningrum, Tenia ;
Khomsah, Siti ;
Suyanto, Suyanto ;
Meliana, Selly ;
Yunanto, Prasti Eko ;
Al Maki, Wikky F. .
2021 4TH INTERNATIONAL SEMINAR ON RESEARCH OF INFORMATION TECHNOLOGY AND INTELLIGENT SYSTEMS (ISRITI 2021), 2020,
[49]   Adaptive Speech Information Hiding Method Based on K-Means [J].
Wu, Zhijun ;
Li, Rong ;
Li, Changliang .
IEEE ACCESS, 2020, 8 :23308-23316
[50]   ADAPTIVE USAGE OF K-MEANS IN EVOLUTIONARY OPTIMIZED DATA CLUSTERING [J].
Wang, Xi ;
Sheng, Weiguo .
PROCEEDINGS OF 2017 INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND CYBERNETICS (ICMLC), VOL 1, 2017, :15-20