Skip to content
Book Open access

Smooth Multi-view Clustering with Anchor and Latent Representation Learning

Nov 2026 · Proceedings of the 34th ACM International Conference on Multimedia · 0 citations · 36 references

Abstract

Tensorized anchor graph-based multi-view clustering (MVC) has shown promising performance on high-dimensional data. However, existing methods generally underexploit latent data structures, while conventional tensor low-rank constraints cannot simultaneously capture intra-view smoothness and inter-view high-order correlations. To address these limitations, we propose SMVCAL, a smooth multi-view clustering framework that jointly performs unified anchor representation learning and latent representation learning. This framework captures underlying structures and alleviates information loss caused by limited anchors. Furthermore, we introduce a tensor total variation Schatten-p regularizer to jointly promote tensor low-rankness and intra-view smoothness, thereby preserving high-order cross-view correlations and local geometric structures. An efficient optimization algorithm is developed for the resulting problem. Extensive experiments on benchmark datasets demonstrate the effectiveness and computational efficiency of SMVCAL. The code is publicly available at: https://github.com/xd-wyp/SMVCAL.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.