DWCL: Dual-Weighted Contrastive Learning for Multi-View Clustering

2024-11-26

arXiv.org

AI Summary

To address unreliable cross-view pairings and representation degradation in multi-view clustering, this paper proposes a dual-weighted contrastive learning framework. First, we introduce the novel Best-Other contrast mechanism, which explicitly models discriminative relationships between the optimal positive sample and suboptimal negative samples. Second, we design a dual-weighting strategy that jointly incorporates view quality and inter-view dissimilarity; we theoretically prove that this strategy mitigates representation degradation and enhances clustering consistency. Our method integrates contrastive learning, multi-view representation learning, weighted loss design, and convergence analysis. Extensive experiments on eight standard multi-view benchmark datasets demonstrate significant improvements over state-of-the-art methods: clustering accuracy increases by 5.4% on Caltech6V7 and 5.6% on MSRCv1. The framework achieves superior robustness and generalizability across diverse multi-view clustering scenarios.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Multi-view contrastive clustering (MVCC) has gained significant attention for generating consistent clustering structures from multiple views through contrastive learning. However, most existing MVCC methods create cross-views by combining any two views, leading to a high volume of unreliable pairs. Furthermore, these approaches often overlook discrepancies in multi-view representations, resulting in representation degeneration. To address these challenges, we introduce a novel model called Dual-Weighted Contrastive Learning (DWCL) for Multi-View Clustering. Specifically, to reduce the impact of unreliable cross-views, we introduce an innovative Best-Other (B-O) contrastive mechanism that enhances the representation of individual views at a low computational cost. Furthermore, we develop a dual weighting strategy that combines a view quality weight, reflecting the quality of each view, with a view discrepancy weight. This approach effectively mitigates representation degeneration by downplaying cross-views that are both low in quality and high in discrepancy. We theoretically validate the efficiency of the B-O contrastive mechanism and the effectiveness of the dual weighting strategy. Extensive experiments demonstrate that DWCL outperforms previous methods across eight multi-view datasets, showcasing superior performance and robustness in MVCC. Specifically, our method achieves absolute accuracy improvements of 5.4% and 5.6% compared to state-of-the-art methods on the Caltech6V7 and MSRCv1 datasets, respectively.

Problem

• Multi-View Clustering
• MVCC Reliability
• View Discrepancy

Innovation

• Dual-weighted Contrastive Learning
• Multi-view Contrastive Clustering
• Best-Other Contrast