Robust Multi-View Spectral Clustering via Low-Rank and Sparse Decomposition

By Rongkai Xia, Yan Pan, Lei Du, Jian Yin. 2014

747 citations recorded by the source corpus.

Multi-view clustering, which seeks a partition of the data inmultiple views that often provide complementary information to eachother, has received considerable attention in recent years. In reallife clustering problems, the data in each view may haveconsiderable noise. However, existing clustering methods blindlycombine the information from multi-view data with possiblyconsiderable noise, which often degrades their performance. In thispaper, we propose a novel Markov chain method for RobustMulti-view Spectral Clustering (RMSC). Our method has a flavor oflow-rank and sparse decomposition, where we firstly construct atransition probability matrix from each single view, and then usethese matrices to recover a shared low-rank transition probabilitymatrix as a crucial input to the standard Markov chain methodfor clustering. The optimization problem of RMSC has a low-rankconstraint on the transition probability matrix, and simultaneouslya probabilistic simplex constraint on each of its rows. To solvethis challenging optimization problem, we propose an optimization procedurebased on the Augmented Lagrangian Multiplier scheme. Experimentalresults on various real world datasets show that theproposed method has superior performance over severalstate-of-the-art methods for multi-view clustering.

Subfields: Artificial Intelligence, Computational Mechanics, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Shirui Pan
2. Dacheng Tao
3. Chengqi Zhang
4. Guodong Long
5. Jing Jiang
6. Bing Liu
7. Ruiqi Hu
8. Hao Wang

In Computational Mechanics

1. Changqing Zhang
2. Huan Fu
3. Xiaochun Cao
4. Dacheng Tao
5. Lei Zhang
6. Zhouchen Lin
7. Hongyuan Zha
8. Jianlong Wu

In Computer Vision and Pattern Recognition

1. Dacheng Tao
2. Luc Van Gool
3. Yi Yang
4. Ming-Hsuan Yang
5. Lei Zhang
6. Ling Shao
7. Xiaogang Wang
8. Andrea Vedaldi

Citations in the corpus

• On Unifying Multi-view Self-Representations for Clustering by Tensor Multi-rank Minimization (2016, methodology): “By using the more powerful deep feature, the baseline algorithms (SPC and LRR) perform better than some sophisticated methods such as RMSC and DiMSC.”
• Consistent and Specific Multi-View Subspace Clustering (2018, background): “Based on a variety of theories, a lot of methods have been developed to extract comprehensive information from multiple views (Xu, Tao, and Xu 2013), including co-training (Blum and Mitchell 1998; Kumar, Rai, and Daumé 2011; Kumar and Daumé 2011), multiple kernel learning (Gönen and Alpaydın 2011) a”
• Multi-view clustering: A survey (2018, neutral): “To address the issue that data may have considerable noise, Xia et al.[81] investigated the Markov chain in order to formulate a multi-view spectral clustering model.”
• Generalized Latent Multi-View Subspace Clustering (2020, background): “2) RMSC [55] recovers a shared low-rank transition probability matrix as the input to the standard Markov chain.”
• Multi-view Contrastive Graph Clustering (2021, positive): “Compared multi-view clustering methods include: PMNE [Liu et al., 2017], RMSC [Xia et al., 2014], SwMC [Nie et al., 2017].”
• Multi-omic and multi-view clustering algorithms: review and cancer benchmark (2018, background): “Xia (27) also examines random walks on the networks, but argues that the stationary distribution in each network can still be noisy.”
• Fast Parameter-Free Multi-View Subspace Clustering With Consensus Anchor Guidance (2021, methodology): “Most MVSC methods adopt multi-view subspace regularization stereotype to capture the inter-view relationships for multi-view clustering [10]– [18].”
• Multiview Spectral Clustering via Structured Low-Rank Matrix Factorization (2017, positive): “To this end, Low-Rank Representation (LRR) [ Xia et al. , 2014; Liu et al. , 2010; Wang et al. , 2015c; Liu and Yan, 2011; Liu et al. , 2013 ] is proposed.”
• MGAE: Marginalized Graph Autoencoder for Graph Clustering (2017, methodology): “, similar to the multi-view representation learning algorithm clustering problems [34, 36], we calculate ) to make sure our representation is symmetric and nonnegative.”
• Learning Graph Embedding With Adversarial Training Methods (2019, methodology): “7) RMSC [63] is a multiview clustering algorithm which”
• Latent Multi-view Subspace Clustering (2017, background): “RMSC [27] recovers a shared low-rank transition probability matrix as input to the standard Markov chain.”
• Graph Learning for Multiview Clustering (2018, neutral): “clustering methods have the same problems, such as methods including multimodal SC [20], robust multiview SC (RMSC) [21], multiview SC (MVSC) [22], SC with brainstorming [23], etc.”
• Essential Tensor Learning for Multi-View Spectral Clustering (2018, methodology): “Considering that multi-view data might be noisy, RMSC [6] hopes to recover a shared low-rank transition probability matrix for the Markov chain based spectral clustering.”
• Multi-view learning overview: Recent progress and new challenges (2017, methodology): “[26] proposed a novel Markov chain method for robust multi-view spectral clustering, which combines the transition probability matrices constructed from each view into a shared transition probability matrix via low-rank and sparse decomposition.”
• GMC: Graph-Based Multi-View Clustering (2020, methodology): “Our work is also related to multi-view spectral clustering [18], [19], [20], [21], [22], [23], [24], [25].”