Large-Scale Multi-View Subspace Clustering in Linear Time

Kang, Zhao; Zhou, Wang-Tao; Zhao, Zhitong; Shao, Jun; Han, Meng; Xu, Zenglin

doi:10.1609/aaai.v34i04.5867

Cited by 262 publications

(93 citation statements)

References 28 publications

(40 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…To the best of our knowledge, few researches have been established to reduce this gap. However, self-representation method, on the other side, benefits from the built coefficient matrix, for it reflects the precise relationships among data entries, making it feasible to achieve a better clustering performance [16]. We also validate this by conducting an ablation study in the experiment part.…”

Section: Missing Data Imputationmentioning

confidence: 77%

Self-Representation Subspace Clustering for Incomplete Multi-view Data

Liu

Zhang

et al. 2021

Proceedings of the 29th ACM International Conference on Multimedia

View full text Add to dashboard Cite

Incomplete multi-view clustering is an important research topic in multimedia where partial data entries of one or more views are missing. Current subspace clustering approaches mostly employ matrix factorization on the observed feature matrices to address this issue. Meanwhile, self-representation technique is left unexplored, since it explicitly relies on full data entries to construct the coefficient matrix, which is contradictory to the incomplete data setting. However, it is widely observed that self-representation subspace method enjoys a better clustering performance over the factorization based one. Therefore, we adapt it to incomplete data by jointly performing data imputation and self-representation learning. To the best of our knowledge, this is the first attempt in incomplete multi-view clustering literature. Besides, the proposed method is carefully compared with current advances in experiment with respect to different missing ratios, verifying its effectiveness. CCS CONCEPTS• Computing methodologies → Cluster analysis; Statistical relational learning; Spectral methods; • Theory of computation → Unsupervised learning and clustering; • Mathematics of computing → Nonconvex optimization.

show abstract

Section: Missing Data Imputationmentioning

confidence: 77%

Self-Representation Subspace Clustering for Incomplete Multi-view Data

Liu

Zhang

et al. 2021

Proceedings of the 29th ACM International Conference on Multimedia

View full text Add to dashboard Cite

show abstract

“…In recent years, many methods are proposed to improve the scalability of SSC. Some scalable SSC solvers [7], [18] suggest to utilize a small subset to represent the entire dataset to reduce the computational cost. The recently proposed Scalable Sparse Subspace Clustering by Orthogonal Matching Pursuit (OMP-SSC) [8] adopts orthogonal matching pursuit to SSC and works well on massive data.…”

Section: B Discussion About Scalable Subspace Clusteringmentioning

confidence: 99%

“…The expensive memory and time consumption severely limits its scalability to stream data in which the size of data is quite huge and even infinite. Although there have been some scalable solvers proposed recently [7]- [11], they cannot tackle the situation under which points in the same clusters may arrive at different times and so the structure of underlying subspaces changes over time. As a result, there is a need to develop a flexible and efficient SSC method for stream data.…”

Section: Introductionmentioning

confidence: 99%

Sparse Subspace Clustering for Stream Data

et al. 2021

View full text Add to dashboard Cite

In the past few years, sparse subspace clustering (SSC) has gained many studies and found wide applications. However, SSC suffers from the limitation in scalability. Furthermore, current SSC methods could hardly tackle stream data where the structure of subspaces may change along time. In this paper, we propose a novel method to extend SSC to stream data (StreamSSC). Our method is based on maintaining a small subset of representatives to characterize the structure of the underlying subspaces during stream data. StreamSSC is efficient in both computation and memory. Experimental results on both synthetic and real-world streams demonstrate the effectiveness of StreamSSC. For efficiency, StreamSSC is faster than existing online subspace clustering methods by roughly a magnitude.

show abstract

“…To address the above challenge, instead of using all samples to reconstructX in (4), we only choose m(m n) representative points [Kang et al, 2020b], i.e., nodes that play an important role in the graph, whose attributes construct…”

Section: Methodsmentioning

confidence: 99%

Graph Filter-based Multi-view Attributed Graph Clustering

Lin

Kang

2021

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

Graph clustering has become an important research topic due to the proliferation of graph data. However, existing methods suffer from two major drawbacks. On the one hand, most methods can not simultaneously exploit attribute and graph structure information. On the other hand, most methods are incapable of handling multi-view data which contain sets of different features and graphs. In this paper, we propose a novel Multi-view Attributed Graph Clustering (MvAGC) method, which is simple yet effective. Firstly, a graph filter is applied to features to obtain a smooth representation without the need of learning the parameters of neural networks. Secondly, a novel strategy is designed to select a few anchor points, so as to reduce the computation complexity. Thirdly, a new regularizer is developed to explore high-order neighborhood information. Our extensive experiments indicate that our method works surprisingly well with respect to state-of-the-art deep neural network methods. The source code is available at https://github.com/sckangz/MvAGC.

show abstract

Large-Scale Multi-View Subspace Clustering in Linear Time

Cited by 262 publications

References 28 publications

Self-Representation Subspace Clustering for Incomplete Multi-view Data

Self-Representation Subspace Clustering for Incomplete Multi-view Data

Sparse Subspace Clustering for Stream Data

Graph Filter-based Multi-view Attributed Graph Clustering

Contact Info

Product

Resources

About