Self-Supervised Deep Multi-View Subspace Clustering

Xiukun Sun; Miaomiao Cheng; Chen Min; Liping Jing

Self-Supervised Deep Multi-View Subspace Clustering

Xiukun Sun, Miaomiao Cheng, Chen Min, Liping Jing ;

Proceedings of The Eleventh Asian Conference on Machine Learning, PMLR 101:1001-1016, 2019.

Abstract

As a new occurring unsupervised method, multi-view clustering offers a good way to investigate the hidden structure from multi-view data and attracts extensive attention in the community of machine learning and data mining. One popular approach is to identify a common latent subspace for capturing the multi-view information. However, these methods are still limited due to the unsupervised learning process and suffer from considerable noisy information from different views. To address this issue, we present a novel multi-view subspace clustering method, named self-supervised deep multi-view subspace clustering (\textbf{S2DMVSC}). It seamlessly integrates spectral clustering and affinity learning into a deep learning framework. \textbf{S2DMVSC} has two main merits. One is that the clustering results can be sufficiently exploited to supervise the latent representation learning for each view (via a classification loss) and the common latent subspace learning (via a spectral clustering loss) for multiple views. The other is that the affinity matrix among data objects is automatically computed according to the high-level and cluster-driven representation. Experiments on two scenarios, including original features and multiple hand-crafted features, demonstrate the superiority of the proposed approach over the state-of-the-art baselines.

Related Material

Download PDF

@InProceedings{pmlr-v101-sun19a, title = {Self-Supervised Deep Multi-View Subspace Clustering}, author = {Sun, Xiukun and Cheng, Miaomiao and Min, Chen and Jing, Liping}, booktitle = {Proceedings of The Eleventh Asian Conference on Machine Learning}, pages = {1001--1016}, year = {2019}, editor = {Lee, Wee Sun and Suzuki, Taiji}, volume = {101}, series = {Proceedings of Machine Learning Research}, address = {Nagoya, Japan}, month = {17--19 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v101/sun19a/sun19a.pdf}, url = {http://proceedings.mlr.press/v101/sun19a.html}, abstract = {As a new occurring unsupervised method, multi-view clustering offers a good way to investigate the hidden structure from multi-view data and attracts extensive attention in the community of machine learning and data mining. One popular approach is to identify a common latent subspace for capturing the multi-view information. However, these methods are still limited due to the unsupervised learning process and suffer from considerable noisy information from different views. To address this issue, we present a novel multi-view subspace clustering method, named self-supervised deep multi-view subspace clustering (\textbf{S2DMVSC}). It seamlessly integrates spectral clustering and affinity learning into a deep learning framework. \textbf{S2DMVSC} has two main merits. One is that the clustering results can be sufficiently exploited to supervise the latent representation learning for each view (via a classification loss) and the common latent subspace learning (via a spectral clustering loss) for multiple views. The other is that the affinity matrix among data objects is automatically computed according to the high-level and cluster-driven representation. Experiments on two scenarios, including original features and multiple hand-crafted features, demonstrate the superiority of the proposed approach over the state-of-the-art baselines.} }

%0 Conference Paper %T Self-Supervised Deep Multi-View Subspace Clustering %A Xiukun Sun %A Miaomiao Cheng %A Chen Min %A Liping Jing %B Proceedings of The Eleventh Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2019 %E Wee Sun Lee %E Taiji Suzuki %F pmlr-v101-sun19a %I PMLR %J Proceedings of Machine Learning Research %P 1001--1016 %U http://proceedings.mlr.press %V 101 %W PMLR %X As a new occurring unsupervised method, multi-view clustering offers a good way to investigate the hidden structure from multi-view data and attracts extensive attention in the community of machine learning and data mining. One popular approach is to identify a common latent subspace for capturing the multi-view information. However, these methods are still limited due to the unsupervised learning process and suffer from considerable noisy information from different views. To address this issue, we present a novel multi-view subspace clustering method, named self-supervised deep multi-view subspace clustering (\textbf{S2DMVSC}). It seamlessly integrates spectral clustering and affinity learning into a deep learning framework. \textbf{S2DMVSC} has two main merits. One is that the clustering results can be sufficiently exploited to supervise the latent representation learning for each view (via a classification loss) and the common latent subspace learning (via a spectral clustering loss) for multiple views. The other is that the affinity matrix among data objects is automatically computed according to the high-level and cluster-driven representation. Experiments on two scenarios, including original features and multiple hand-crafted features, demonstrate the superiority of the proposed approach over the state-of-the-art baselines.