Dual Shared-Specific Multiview Subspace Clustering

Zhou, Tao; Zhang, Changqing; Peng, Xi; Bhaskar, Harish; Yang, Jie

doi:10.1109/tcyb.2019.2918495

Cited by 132 publications

(28 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Among them, Z = {z n } N n=1 ∈ R M h ×N is the joint latent representation fused by view-specific and consistent representations according to equation (7) with the contribution weights Figure 1: AMVDSN's Framework: Take data with two views as an example. X 1 and X 2 (view 1 and 2) are embedded into the same dimension through Encoder 1 and Encoder 2 with structure of shortcut connection, and the corresponding latent representations are H 1 and H 2 respectively.…”

Section: Framework Descriptionmentioning

confidence: 99%

“…In this paper, we apply shortcut connection in the modules that have more than two layers, i.e., encoder, decoder, attentive layers, and therefore we should reformulate equation ( 2), (7), and ( 13) respectively as follows:…”

Section: Framework Descriptionmentioning

confidence: 99%

“…Most existing so called multi-view algorithms are belong to multi-feature learning [1], and this paper is also under the assumption of multi-feature data. On the other hand, in this field, multi-view self-representation subspace learning is one of the most attractive directions, where most existing multi-view subspace clustering algorithms are actually self-representation based methods [2,3,4,5,6,7].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Attentive multi-view deep subspace clustering net

Liu

Zuo

2021

Neurocomputing

View full text Add to dashboard Cite

Section: Framework Descriptionmentioning

confidence: 99%

Section: Framework Descriptionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Attentive multi-view deep subspace clustering net

Liu

Zuo

2021

Neurocomputing

View full text Add to dashboard Cite

“…The goal of multiview subspace clustering is to simultaneously cluster data into respective clusters using multiview feature sets, where each cluster corresponds to a lowdimensional subspace [45], [53], [57], [60], [63].Multiview data consist of different views and can be represented by multiple distinct feature sets. The consistency of different feature sets indicates that common information is shared among different views.…”

Section: Introductionmentioning

confidence: 99%

Multiview Subspace Clustering Using Low-Rank Representation

Chen

Yang

Mao

et al. 2022

IEEE Trans. Cybern.

View full text Add to dashboard Cite

Multiview subspace clustering is one of the most widely used methods for exploiting the internal structures of multiview data. Most previous studies have performed the task of learning multiview representations by individually constructing an affinity matrix for each view without simultaneously exploiting the intrinsic characteristics of multiview data. In this paper, we propose a multiview low-rank representation (MLRR) method to comprehensively discover the correlation of multiview data for multiview subspace clustering. MLRR considers symmetric low-rank representations (LRRs) to be an approximately linear spatial transformation under the new base, i.e., the multiview data themselves, to fully exploit the angular information of the principal directions of LRRs, which is adopted to construct an affinity matrix for multiview subspace clustering, under a symmetric condition. MLRR takes full advantage of LRR techniques and a diversity regularization term to exploit the diversity and consistency of multiple views, respectively, and this method simultaneously imposes a symmetry constraint on LRRs. Hence, the angular information of the principal directions of rows is consistent with that of columns in symmetric LRRs. The MLRR model can be efficiently calculated by solving a convex optimization problem. Moreover, we present an intuitive fusion strategy for symmetric LRRs from the perspective of spectral clustering to obtain a compact representation, which can be shared by multiple views and comprehensively represents the intrinsic features of multiview data. Finally, the experimental results based on benchmark datasets demonstrate the effectiveness and robustness of MLRR compared with several state-of-the-art multiview subspace clustering algorithms.

show abstract

“…Therefore, multi-view learning could leverage both consistency and complementarity to explore more compact and complete information from multiple observations. Subspace clustering is one of the most attractive research topics in the field of multi-view learning, especially for algorithms based on self-expressiveness [6,7,8,9,10,11,12,13,14], which state each data point in a union of subspaces can be represented by a linear combination of all other data points [15]. Existing multi-view subspace clustering usually focus on two perspectives: the prior constraint on subspace representation and the construction method of latent representation.…”

Section: Introductionmentioning

confidence: 99%

Partially latent factors based multi-view subspace learning

Lu¹,

Liu²,

Liu³

et al. 2022

Preprint

View full text Add to dashboard Cite

Multi-view subspace clustering always performs well in high-dimensional data analysis, but is sensitive to the quality of data representation. To this end, a two stage fusion strategy is proposed to embed representation learning into the process of multi-view subspace clustering. This paper first propose a novel matrix factorization method that can separate the coupling consistent and complementary information from observations of multiple views. Based on the obtained latent representations, we further propose two subspace clustering strategies: feature-level fusion and subspace-level hierarchical strategy. Feature-level method concatenates all kinds of latent representations from multiple views, and the original problem therefore degenerates to a single-view subspace clustering process. Subspace-level hierarchical method performs different self-expressive reconstruction processes on the corresponding complementary and consistent latent representations coming from each view, i.e. the prior constraints imposed on different types of subspace representations are related to the appropriate input factors. Finally, extensive experimental results on real-world datasets demonstrate the superiority of our proposed methods by comparing against some state-of-the-art subspace clustering algorithms.

show abstract

Dual Shared-Specific Multiview Subspace Clustering

Cited by 132 publications

References 46 publications

Attentive multi-view deep subspace clustering net

Attentive multi-view deep subspace clustering net

Multiview Subspace Clustering Using Low-Rank Representation

Partially latent factors based multi-view subspace learning

Contact Info

Product

Resources

About