Tensorized Anchor Graph Learning for Large-scale Multi-view Clustering (2024)

Dai, J., Ren, Z., Luo, Y., Song, H. (2023). Tensorized Anchor Graph Learning for Large-scale Multi-view Clustering. Cognitive Computation, 15(5), 1581-1592. https://doi.org/10.1007/s12559-023-10146-3

Dai, Jian ; Ren, Zhenwen ; Luo, Yunzhi 等. / Tensorized Anchor Graph Learning for Large-scale Multi-view Clustering. 在: Cognitive Computation. 2023 ; 卷 15, 号码 5. 页码 1581-1592.

@article{8b99b8229d6f483aa365f6f3cfe235ca,

title = "Tensorized Anchor Graph Learning for Large-scale Multi-view Clustering",

abstract = "With the continuous development of information acquisition technologies, large-scale multi-view data increases rapidly. The enormous computational and storage complexity makes it very challenging to process these data in real-world applications. Most existing multi-view subspace clustering (MVSC) always suffers from quadratic space complexity and quadratic or even cubic time complexity, resulting in extreme limitations for large-scale tasks. Meanwhile, the original data usually contain lots of noise or redundant features, which further enhances the difficulty of the large-scale clustering tasks. This paper proposes a novel MVSC method for efficiently and effectively dealing with large-scale multi-view data, termed as tensorized anchor graph learning (TAGL) for large-scale multi-view clustering. Concretely, TAGL first projects the original multi-view data from the original space into the latent embedding space, where the view-consistent anchor matrix. Meanwhile, we establish the connection between the anchor matrix and the original data to construct multiple view-specific anchor graphs. Furthermore, these anchor graphs are stacked into a graph tensor to capture the high-order correlation. Finally, by developing an effective optimization algorithm, the high-quality anchors, anchor graph, and anchor graph tensor can be jointly learned in a mutually reinforcing way. Experimental results on several big sizes of datasets verify the superiority and validity of TAGL. Therefore, the proposed TAGL can efficiently and effectively handle large-scale data tasks for real-world applications.",

keywords = "Anchor graph learning, Low-rank tensor, Multi-view subspace clustering, Redundant features removal",

author = "Jian Dai and Zhenwen Ren and Yunzhi Luo and Hong Song and Jian Yang",

note = "Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.",

year = "2023",

month = sep,

doi = "10.1007/s12559-023-10146-3",

language = "English",

volume = "15",

pages = "1581--1592",

journal = "Cognitive Computation",

issn = "1866-9956",

publisher = "Springer New York",

number = "5",

}

Dai, J, Ren, Z, Luo, Y, Song, H 2023, 'Tensorized Anchor Graph Learning for Large-scale Multi-view Clustering', Cognitive Computation, 卷 15, 号码 5, 页码 1581-1592. https://doi.org/10.1007/s12559-023-10146-3

Tensorized Anchor Graph Learning for Large-scale Multi-view Clustering. / Dai, Jian; Ren, Zhenwen; Luo, Yunzhi 等.
在: Cognitive Computation, 卷 15, 号码 5, 09.2023, 页码 1581-1592.

科研成果: 期刊稿件 › 文章 › 同行评审

TY - JOUR

T1 - Tensorized Anchor Graph Learning for Large-scale Multi-view Clustering

AU - Dai, Jian

AU - Ren, Zhenwen

AU - Luo, Yunzhi

AU - Song, Hong

AU - Yang, Jian

N1 - Publisher Copyright:© 2023, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.

PY - 2023/9

Y1 - 2023/9

N2 - With the continuous development of information acquisition technologies, large-scale multi-view data increases rapidly. The enormous computational and storage complexity makes it very challenging to process these data in real-world applications. Most existing multi-view subspace clustering (MVSC) always suffers from quadratic space complexity and quadratic or even cubic time complexity, resulting in extreme limitations for large-scale tasks. Meanwhile, the original data usually contain lots of noise or redundant features, which further enhances the difficulty of the large-scale clustering tasks. This paper proposes a novel MVSC method for efficiently and effectively dealing with large-scale multi-view data, termed as tensorized anchor graph learning (TAGL) for large-scale multi-view clustering. Concretely, TAGL first projects the original multi-view data from the original space into the latent embedding space, where the view-consistent anchor matrix. Meanwhile, we establish the connection between the anchor matrix and the original data to construct multiple view-specific anchor graphs. Furthermore, these anchor graphs are stacked into a graph tensor to capture the high-order correlation. Finally, by developing an effective optimization algorithm, the high-quality anchors, anchor graph, and anchor graph tensor can be jointly learned in a mutually reinforcing way. Experimental results on several big sizes of datasets verify the superiority and validity of TAGL. Therefore, the proposed TAGL can efficiently and effectively handle large-scale data tasks for real-world applications.

AB - With the continuous development of information acquisition technologies, large-scale multi-view data increases rapidly. The enormous computational and storage complexity makes it very challenging to process these data in real-world applications. Most existing multi-view subspace clustering (MVSC) always suffers from quadratic space complexity and quadratic or even cubic time complexity, resulting in extreme limitations for large-scale tasks. Meanwhile, the original data usually contain lots of noise or redundant features, which further enhances the difficulty of the large-scale clustering tasks. This paper proposes a novel MVSC method for efficiently and effectively dealing with large-scale multi-view data, termed as tensorized anchor graph learning (TAGL) for large-scale multi-view clustering. Concretely, TAGL first projects the original multi-view data from the original space into the latent embedding space, where the view-consistent anchor matrix. Meanwhile, we establish the connection between the anchor matrix and the original data to construct multiple view-specific anchor graphs. Furthermore, these anchor graphs are stacked into a graph tensor to capture the high-order correlation. Finally, by developing an effective optimization algorithm, the high-quality anchors, anchor graph, and anchor graph tensor can be jointly learned in a mutually reinforcing way. Experimental results on several big sizes of datasets verify the superiority and validity of TAGL. Therefore, the proposed TAGL can efficiently and effectively handle large-scale data tasks for real-world applications.

KW - Anchor graph learning

KW - Low-rank tensor

KW - Multi-view subspace clustering

KW - Redundant features removal

UR - http://www.scopus.com/inward/record.url?scp=85158082660&partnerID=8YFLogxK

U2 - 10.1007/s12559-023-10146-3

DO - 10.1007/s12559-023-10146-3

M3 - Article

AN - SCOPUS:85158082660

SN - 1866-9956

VL - 15

SP - 1581

EP - 1592

JO - Cognitive Computation

JF - Cognitive Computation

IS - 5

ER -

Dai J, Ren Z, Luo Y, Song H, Yang J. Tensorized Anchor Graph Learning for Large-scale Multi-view Clustering. Cognitive Computation. 2023 9月;15(5):1581-1592. doi: 10.1007/s12559-023-10146-3