Multiview Clustering of Images with Tensor Rank Minimization via Nonconvex Approach

Yang, Ming; Luo, Qilun; Wen, Li; Xiao, Mingqing

doi:10.1137/20m1318006

Cited by 15 publications

(4 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Equation ( 3) can be solved by applying a norm threshold action on singular values of the observation matrix in the form of Equation ( 4) [11,12]:…”

Section: Proposed Methodsmentioning

confidence: 99%

Prediction of Drug-Target Protein Interaction Based on the Minimization of Weighted Nuclear Norm and Similarity Graph between Drugs and Target Proteins

Sorkhi

Hashemi

Iranpour

et al. 2021

IJE

View full text Add to dashboard Cite

Identification of drug-target protein interaction plays an important role in the drug discovery process. Given the fact that prediction experiments are time-consuming, tedious, and very costly, the computational prediction could be a proper solution for decreasing search space for evaluation of the interaction between drug and target. In this paper, a novel approach based on the known drug-target interactions based on similarity graphs is proposed. It was shown that use of this method was a lowranking issue and WNNM (weighted nuclear norm minimization) method was applied to detect the drug-target interactions. In the proposed method, the interaction between the drug and the target is encoded by graphs. Also known drug-target interaction, drug-drug similarity, target-target and combination of similarities were used as input. The proposed method was performed on four benchmark datasets, including enzymes (Es), ion channels (IC), G protein-coupled receptors (GPCRs), and nuclear receptors (NRs) based on the AUC and AUPR criteria. Finally, the results showed the improved performance of the proposed method.

show abstract

“…Equation ( 3) can be solved by applying a norm threshold action on singular values of the observation matrix in the form of Equation ( 4) [11,12]:…”

Section: Proposed Methodsmentioning

confidence: 99%

Prediction of Drug-Target Protein Interaction Based on the Minimization of Weighted Nuclear Norm and Similarity Graph between Drugs and Target Proteins

Sorkhi

Hashemi

Iranpour

et al. 2021

IJE

View full text Add to dashboard Cite

show abstract

“…Here we illustrate the advantage of using nonconvex MCP, compared with the nuclear norm and the log-determinant approximation rank function used in the literature [39]. The log-determinant approximation rank function for X ∈ R M×N is given by…”

Section: Motivationmentioning

confidence: 99%

“…* Extended YaleB 4 contains 38 individuals and around 64 near frontal images under different illuminations for each individual. Similar to other works [39,43], we use images for the first 10 classes, including 640 frontal face images. Also, the intensity, LBP and Gabor features are extracted for clustering.…”

Section: Datasets Competitors and Evaluation Measuresmentioning

confidence: 99%

“…For tensor case, although the tensor structure is much more complicated than the matrix case in terms of spatial structures, there also exists the nonconvex tensor rank approximation method for multi-view subspace clustering problem. For example, Yang et al [39] recently proposed a tensor log-determinant (TLD) function as a tighter rank approximation to reduce the gap among the singular values so as to retain the main information of different subspaces under the nonconvex framework. This approach provides quite desirable results in clustering but it only considers the global low-rank structure of the sample space and does not address the inherent non-linearity of the feature spaces, which usually occur in certain complex datasets, resulting in unsatisfactory outcomes in such cases.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Nonconvex multi-view subspace clustering via simultaneously learning the representation tensor and affinity matrix*

Li¹,

Li²,

Xiao³

2022

Inverse Problems

Self Cite

View full text Add to dashboard Cite

Multi-view subspace clustering, which aims to partition a dataset into its relevant subspaces based on their multi-view features, has been widely applied to identify various characteristics of datasets. The typical model of multi-view subspace clustering in literature often makes use of the nuclear norm to seek the underlying low-rank representation. However, due to the sum property of the singular values deﬁned by tensor nuclear norm, the existing multi-view subspace clustering does not well handle the noise and the illumination variations embedded in multi-view data. To address and improve the robustness and clustering performance, we propose a new nonconvex multi-view subspace clustering model via tensor minimax concave penalty (MCP) approximation associated with rank minimization (NMSC-MCP), which can simultaneously construct the low-rank representation tensor and aﬃnity matrix in a uniﬁed framework. Speciﬁcally, the nonconvex MCP approximation rank function is adopted to as a tighter tensor rank approximation to discriminate the dimension of features so that better accuracy can be achieved. In addition, we also address the local structure by including both hyper-Laplacian regularization and auto-weighting scheme into the objective function to promote the clustering performance. A corresponding iterative algorithm is then developed to solve the proposed model and the constructed iterative sequence generated by the proposed algorithm is shown to converge to the desirable KKT critical point. Extensive experiments on benchmark datasets have demonstrate the highly desirable eﬀectiveness of our proposed method.

show abstract

Proximal gradient algorithm for nonconvex low tubal rank tensor recovery

2023

View full text Add to dashboard Cite

Multiview Clustering of Images with Tensor Rank Minimization via Nonconvex Approach

Cited by 15 publications

References 33 publications

Prediction of Drug-Target Protein Interaction Based on the Minimization of Weighted Nuclear Norm and Similarity Graph between Drugs and Target Proteins

Prediction of Drug-Target Protein Interaction Based on the Minimization of Weighted Nuclear Norm and Similarity Graph between Drugs and Target Proteins

Nonconvex multi-view subspace clustering via simultaneously learning the representation tensor and affinity matrix*

Proximal gradient algorithm for nonconvex low tubal rank tensor recovery

Contact Info

Product

Resources

About