Discriminative Anchor Learning for Efficient Multi-view Clustering

TL;DR

The paper tackles the efficiency and effectiveness of multi-view clustering by identifying limitations of fixed-shared anchor graphs and proposing discriminative anchor learning (DALMC). It jointly learns discriminative view-specific representations and a consensus anchor graph under orthogonal constraints, enabling linear-time complexity and improved cross-view complementarity. The method demonstrates superior clustering performance and scalability across multiple large datasets, supported by thorough ablations, parameter analyses, and convergence studies. This approach offers a practical and scalable solution for large-scale multi-view data analysis with enhanced representation quality.

Abstract

Multi-view clustering aims to study the complementary information across views and discover the underlying structure. For solving the relatively high computational cost for the existing approaches, works based on anchor have been presented recently. Even with acceptable clustering performance, these methods tend to map the original representation from multiple views into a fixed shared graph based on the original dataset. However, most studies ignore the discriminative property of the learned anchors, which ruin the representation capability of the built model. Moreover, the complementary information among anchors across views is neglected to be ensured by simply learning the shared anchor graph without considering the quality of view-specific anchors. In this paper, we propose discriminative anchor learning for multi-view clustering (DALMC) for handling the above issues. We learn discriminative view-specific feature representations according to the original dataset and build anchors from different views based on these representations, which increase the quality of the shared anchor graph. The discriminative feature learning and consensus anchor graph construction are integrated into a unified framework to improve each other for realizing the refinement. The optimal anchors from multiple views and the consensus anchor graph are learned with the orthogonal constraints. We give an iterative algorithm to deal with the formulated problem. Extensive experiments on different datasets show the effectiveness and efficiency of our method compared with other methods.

Figures (6)

• Figure 1: (a) Overview of our framework. We first learn discriminative feature representations of the original multi-view dataset, then the consensus anchor graph is built on these feature representations. Lastly, with the obtained consensus anchor graph, $K$-means is employed to achieve the final clustering results.
• Figure 2: Parameter investigation of $\beta$ on all datasets in terms of ACC and NMI.
• Figure 3: Parameter investigation of $\beta$ on all datasets in terms of F1-score and Purity.
• Figure 4: Quantitative study of anchors on all datasets in terms of ACC and NMI.
• Figure 5: Quantitative study of anchors on all datasets in terms of F1-score and Purity.