Robust Multi-view Clustering against Imperfect Information

2026-06-02

AI Summary

Real-world multi-view data often suffer simultaneously from incomplete views (IV) and cross-view sample non-correspondence (NC), yet existing methods struggle to handle both issues in a unified framework. This work proposes a probabilistic latent correspondence inference framework, PLCI, which attributes IV and NC to imperfect cross-view correspondence information. By modeling the correspondence as a latent variable, PLCI jointly leverages instance-level reliability estimation and prototype-level semantic transfer to infer its posterior distribution in an end-to-end manner, thereby enabling robust clustering. Extensive experiments on six benchmark datasets demonstrate that PLCI significantly outperforms ten state-of-the-art methods, confirming its effectiveness and superiority in handling imperfect multi-view scenarios.

Abstract

Real-world multi-view data always suffer from imperfect information problem, where the view-specific observations are absent (i.e., Incomplete Views, IV) and cross-view correspondences are mismatched (i.e., Noisy Correspondences, NC) for certain instances. As a remedy, numerous IV- and NC-oriented multi-view clustering (MvC) methods have been proposed, which however require either reliable correspondences or sufficiently complete instances, thus stopping short of addressing the imperfect information problem. In contrast, we observe that both IV and NC challenges originate from the same issue of imperfect cross-view counterpart information, where the counterpart of an anchor instance in another view might be either unavailable or unreliable. Based on the observation, we propose a novel robust MvC framework, termed Posterior-guided Latent Counterpart Inference (PLCI), which could handle both IV and NC in a unified manner. Specifically, PLCI formulates the desired cross-view counterpart of each anchor instance as a latent variable, and integrates both instance-level reliability and prototype-level semantic transport to infer the posterior distribution of the latent counterpart. Extensive experiments on six widely-used multi-view datasets against 10 state-of-the-art MvC methods demonstrate the effectiveness of PLCI for tackling the imperfect information problem. The code will be released upon acceptance.

Problem

• Incomplete Views
• Noisy Correspondences
• Multi-view Clustering
• Imperfect Information
• Cross-view Correspondence

Innovation

• multi-view clustering
• incomplete views
• noisy correspondences
• latent counterpart inference
• posterior-guided learning