Deep Safe Multi-view Clustering: Reducing the Risk of Clustering Performance Degradation Caused by View Increase

Tang, Huayi; Liu, Yong

doi:10.1109/cvpr52688.2022.00030

Cited by 33 publications

(15 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Lin et al [27] train a projector in a supervised fashion to predict the unobserved representations. Tang et al [29] and Yang et al [30] fill the unobserved representations with the average of adjacent cross-view features. Although some promising results have been achieved by these studies, almost all of them still heavily rely on paired samples to learn the shared representation or to impute the unobserved representations.…”

Section: Multi-view Clusteringmentioning

confidence: 99%

“…In contrast, some methods aim to learn a shared representation across all views without explicitly imputing the unobserved samples or representations [20], [21], [22], [23], [24], [25], [26]. To capture high nonlinearity, some approaches adopt deep neural networks to predict the representations of unobserved samples, embracing powerful learning and nonlinear modeling abilities [27], [28], [29], [30]. Despite their promising performance, these methods still heavily rely on some paired samples (i.e., both samples are observed and correspond correctly to each other), which are often unavailable in real-world applications.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Robust Multi-View Clustering With Incomplete Information

Yang

et al. 2023

IEEE Trans. Pattern Anal. Mach. Intell.

180

View full text Add to dashboard Cite

Robust multi-view learning with incomplete information has received significant attention due to issues such as incomplete correspondences and incomplete instances that commonly affect real-world multi-view applications. Existing approaches heavily rely on paired samples to realign or impute defective ones, but such preconditions cannot always be satisfied in practice due to the complexity of data collection and transmission. To address this problem, we present a novel framework called SeMantic Invariance LEarning (SMILE) for multi-view clustering with incomplete information that does not require any paired samples. To be specific, we discover the existence of invariant semantic distribution across different views, which enables SMILE to alleviate the cross-view discrepancy to learn consensus semantics without requiring any paired samples. The resulting consensus semantics remains unaffected by cross-view distribution shifts, making them useful for realigning/imputing defective instances and forming clusters. We demonstrate the effectiveness of SMILE through extensive comparison experiments with 13 state-of-the-art baselines on five benchmarks. Our approach improves the clustering accuracy of NoisyMNIST from 19.3%/23.2% to 82.7%/69.0% when the correspondences/instances are fully incomplete. We will release the code after acceptance.

show abstract

Section: Multi-view Clusteringmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Robust Multi-View Clustering With Incomplete Information

Yang

et al. 2023

IEEE Trans. Pattern Anal. Mach. Intell.

180

View full text Add to dashboard Cite

show abstract

“…DCP [11] learns consistent representations and complements information by maximizing mutual information and minimizing cross-entropy. DSIMVC [23] dynamically complements the missing views from the learned semantic neighbors, solving the multi-view missing problem. Implementation.…”

Section: Experimental Settingmentioning

confidence: 99%

Deep Incomplete Multi-view Clustering with Cross-view Partial Sample and Prototype Alignment

Jin¹,

Wang²,

Dong³

et al. 2023

Preprint

View full text Add to dashboard Cite

The success of existing multi-view clustering relies on the assumption of sample integrity across multiple views. However, in real-world scenarios, samples of multi-view are partially available due to data corruption or sensor failure, which leads to incomplete multi-view clustering study (IMVC). Although several attempts have been proposed to address IMVC, they suffer from the following drawbacks: i) Existing methods mainly adopt cross-view contrastive learning forcing the representations of each sample across views to be exactly the same, which might ignore view discrepancy and flexibility in representations; ii) Due to the absence of non-observed samples across multiple views, the obtained prototypes of clusters might be unaligned and biased, leading to incorrect fusion. To address the above issues, we propose a Cross-view Partial Sample and Prototype Alignment Network (CPSPAN) for Deep Incomplete Multi-view Clustering. Firstly, unlike existing contrastive-based methods, we adopt pair-observed data alignment as 'proxy supervised signals' to guide instance-to-instance correspondence construction among views. Then, regarding of the shifted prototypes in IMVC, we further propose a prototype alignment module to achieve incomplete distribution calibration across views. Extensive experimental results showcase the effectiveness of our proposed modules, attaining noteworthy performance improvements when compared to existing IMVC competitors on benchmark datasets.

show abstract

“…However, in this approach, the neural network is mainly used to extract deep low-dimensional features of the image and is not responsible for performing clustering analysis. Tang et al [19] propose a generic framework based on deep learning, where the model is trained to address the risk of performance degradation that may result from increased multiview clustering views by automatically selecting features to simultaneously extract complementary information and discard meaningless noise. Ronen et al [20] proposed a method for deep clustering which allows for non-parametric deep clustering by inferring the values of the parameters from learning.…”

Section: Neural Network Participatory Clusteringmentioning

confidence: 99%

End-to-end unsupervised clustering neural networks for image clustering

Li¹,

Li²,

Zhu³

2023

Preprint

View full text Add to dashboard Cite

<p>In this paper, we propose a new clustering module that can be trained jointly with existing neural network layers. Specifically, we have designed a generic clustering module with a competitive update mechanism. The module consists of a Gaussian unit and a maximum pooling layer. The Gaussian unit forward propagation conforms to the joint Gaussian distribution and contains two sets of trainable parameters. It requires no tedious setup and has a plug-and-play feature. To improve the representation capability of the network, we used an auto-encoder to extract the hidden semantics of the input features and combined the clustering module with the auto-encoder to construct an end-to-end unsupervised clustering neural network. We refer to this as HGL_CAE(High-dimensional Gaussian distribution layers combined with convolutional autoencoders).The network is highly adaptable to different input feature dimensions and can cope with situations where the number of clusters cannot be determined in advance. We conducted experiments on the MNIST dataset and the Fashion_MNIST dataset with a clustering accuracy of 93.38% and 72.83% respectively. It is highly competitive with existing methods. </p>

show abstract

Deep Safe Multi-view Clustering: Reducing the Risk of Clustering Performance Degradation Caused by View Increase

Cited by 33 publications

References 31 publications

Robust Multi-View Clustering With Incomplete Information

Robust Multi-View Clustering With Incomplete Information

Deep Incomplete Multi-view Clustering with Cross-view Partial Sample and Prototype Alignment

End-to-end unsupervised clustering neural networks for image clustering

Contact Info

Product

Resources

About