Fast and Scalable Incomplete Multi-View Clustering with Duality Optimal Graph Filtering

Liang Du, Yukai Shi, Yan Chen, Peng Zhou, Yuhua Qian

Abstract: Incomplete Multi-View Clustering (IMVC) is crucial for multi-media data analysis. While graph learning-based IMVC methods have shown promise, they still have limitations. The prevalent first-order affinity graph often misclassifies out-neighborhood intra-cluster and in-neighbor inter-cluster samples, worsened by data incompleteness. These inaccuracies, combined with high computational demands, restrict their suitability for large-scale IMVC tasks. To address these issues, we propose a novel Fast and Scalable IMVC with duality Optimal graph Filtering (FSIMVC-OF). Specifically, we refine the clustering-friendly structure of the bipartite graph by learning an optimal filter within a consensus clustering framework. Instead of learning a sample-side filter, we optimize an anchor-side graph filter and apply it to the anchor side, ensuring computational efficiency with linear complexity, supported by the provable equivalence between these two types of graph filters. We present an alternative optimization algorithm with linear complexity. Extensive experimental analysis demonstrates the superior performance of FSIMVC-OF over current IMVC methods. The codes of this article are released in https://github.com/sroytik/FSIMVC-OF.

Primary Subject Area: [Content] Multimodal Fusion

Secondary Subject Area: [Content] Multimodal Fusion

Relevance To Conference:

This research presents a novel method, Fast and Scalable Incomplete Multi-View Clustering with duality Optimal graph Filtering (FSIMVC-OF), which significantly enhances the field of multimedia/multimodal processing. It tackles the challenge of Incomplete Multi-View Clustering (IMVC) by effectively utilizing adaptive higher-order graph filters to reveal complex data relationships in large-scale multi-view datasets.

The FSIMVC-OF method goes beyond traditional approaches that often miss higher-order structures in data due to their reliance on preconstructed incomplete graphs. By integrating consensus graph learning with these advanced graph filters within a unified framework, FSIMVC-OF provides a more nuanced and accurate representation of the multi-view data, leading to improved clustering performance.

Notably, FSIMVC-OF optimizes anchor-side graph filters, ensuring computational efficiency with linear complexity. A key innovation is the development of a linearly complex optimization algorithm, enabling efficient handling of large datasets. Extensive experiments validate FSIMVC-OF's superior performance over existing IMVC methods, setting a new benchmark in multimedia data analysis.

This scalable and robust approach is pivotal for understanding and managing complex multi-view data, advancing multimedia applications requiring multi-perspective data processing.

Submission Number: 3502