Multi-View Clustering via Deep Matrix Factorization
By Handong Zhao, Zhengming Ding, Yun Raymond Fu. AAAI Conference on Artificial Intelligence, 2017
522 citations recorded by the source corpus.
Multi-View Clustering (MVC) has garnered more attention recently since many real-world data are comprised of different representations or views. The key is to explore complementary information to benefit the clustering problem. In this paper, we present a deep matrix factorization framework for MVC, where semi-nonnegative matrix factorization is adopted to learn the hierarchical semantics of multi-view data in a layer-wise fashion. To maximize the mutual information from each view, we enforce the non-negative representation of each view in the final layer to be the same. Furthermore, to respect the intrinsic geometric structure in each view data, graph regularizers are introduced to couple the output representation of deep structures. As a non-trivial contribution, we provide the solution based on alternating minimization strategy, followed by a theoretical proof of convergence. The superior experimental results on three face benchmarks show the effectiveness of the proposed deep matrix factorization model.
Subfields: Computer Vision and Pattern Recognition
View Paper: https://www.semanticscholar.org/paper/f58ee95c2c4bdb1432e15d981dcbdb2038a55184
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computer Vision and Pattern Recognition
Citations in the corpus
- Multi-view clustering: A survey (2018, unknown): “Zhao et al.[90] discussed the semi-supervised MvC, and presented a multi-view matrix completion method with a pairwise similarity matrix in order to utilize side information; namely, must-link and cannot-link.”
- Consistent and Specific Multi-View Subspace Clustering (2018, background): “Deep non-negative matrix matrix factorization (Zhao, Ding, and Fu 2017) builds a deep structure to seek a common feature representation with more consistent knowledge to facilitate clustering.”
- Adversarial Feature Hallucination Networks for Few-Shot Learning (2020, background): “The rich and accessible labeled data fuel the revolutionary success of deep learning [7, 46, 20].”
- Incomplete Multiview Spectral Clustering With Adaptive Graph Learning (2020, methodology): “These features acquired from different views are regarded as the multiview features [4], [5].”
- Doubly Aligned Incomplete Multi-view Clustering (2018, background): “In multi-view data, these different views often share some consistency and complementary information [Sun, 2013; Zhao et al. , 2017].”
- COMIC: Multi-view Clustering Without Parameter Selection (2019, background): “In recent, Andrew et al. (2013); Wang et al. (2015); Zhao et al. (2017); Peng et al. (2017b) have devoted to neural network based MvC by deeply learning a shared representation across different views.”
- Deep Domain Generalization With Structured Low-Rank Constraint (2018, background): “…and increased great attention in a lot of real-world applications, due to its promising ability in different learning tasks, e.g., face recognition [32], image super-resolution [25], action classification [33], [34], feature selection [35], image clustering [36] and multi-view learning [37].”
- Deep Partial Multi-View Learning (2020, methodology): “(7) DMF-MVC (Deep Semi-NMF for Multi-View Clustering) [58] utilizes a deep structure through semi-nonnegative matrix factorization to seek a common feature representation.”
- GMC: Graph-Based Multi-View Clustering (2020, methodology): “The other related multi-view clustering methods can be roughly classified into three categories: co-training style clustering [19], [26], [27], [28], multi-kernel clustering [29], [30], [31], [32], and multi-view subspace clustering [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43].”
- Deep multi-view learning methods: A review (2021, methodology): “[121] extend the deep matrix factorization technique to a multi-view 395 clustering (MVC) method.”
- Multi-level Feature Learning for Contrastive Multi-view Clustering (2021, unknown): “The second category of MVC methods is based on the matrix factorization technique [23,56] that is formally equivalent to the relaxation of K -means [26].”
- A Survey on Multi-View Clustering (2017, unknown): “At the same time, MVC papers appear largely in many top venues like the conferences ICML [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], NIPS [19], [20], CVPR [21], [22], [23], [24], ICCV [25], AAAI [26], [27], [28], [29], [30], IJCAI [31], [32], [33], [34], [35], [36]”