Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

contrastive multi-view clustering

Contrastive multi-view clustering is an unsupervised machine learning paradigm that groups unlabeled data points represented across multiple diverse feature sets or modalities by integrating multi-view learning with contrastive representation learning. In this framework, neural networks project data from different views into shared or coordinated embedding spaces, optimizing objectives that pull representations of the same instance or semantically related samples closer together while pushing distinct or non-related instances apart. By capturing view-invariant consistency alongside complementary cross-modal information, contrastive multi-view clustering produces discriminative feature representations that enable the effective discovery and partitioning of natural cluster structures across complex, multi-source datasets.

1 item

Decoupled Contrastive Multi-View Clustering with High-Order Random Walks

Decoupled Contrastive Multi-View Clustering with High-Order Random Walks

Yiding Lu, Yijie Lin, Mouxing Yang, Dezhong Peng, Peng Hu, Xi Peng

OrganizationsSichuan University

Why you should read this

Proposes a decoupled multi-view clustering framework that uses high-order random walks to rectify false positive and false negative pairs globally while preserving view-specific information through cross-view reconstruction.

In recent, some robust contrastive multi-view clustering (MvC) methods have been proposed, which construct data pairs from neighborhoods to alleviate the false negative issue, i.e., some intra-cluster samples are wrongly treated as negative pairs. Although promising performance has been achieved by these methods, the false negative issue is still far from addressed and the false positive issue emerges because all in- and out-of-neighborhood samples are simply treated as positive and negative, respectively. To address the issues, we propose a novel robust method, dubbed decoupled contrastive multi-view clustering with high-order random walks (DIVIDE). In brief, DIVIDE leverages random walks to progressively identify data pairs in a global instead of local manner. As a result, DIVIDE could identify in-neighborhood negatives and out-of-neighborhood positives. Moreover, DIVIDE embraces a novel MvC architecture to perform inter- and intra-view contrastive learning in different embedding spaces, thus boosting clustering performance and embracing the robustness against missing views. To verify the efficacy of DIVIDE, we carry out extensive experiments on four benchmark datasets comparing with nine state-of-the-art MvC methods in both complete and incomplete MvC settings. The code is released on https://github.com/XLearning-SCU/2024-AAAI-DIVIDE.

Added

2026-09-26