Multi-View Information-Bottleneck Representation Learning

In real-world applications, clustering or classification can usually be improved by fusing information from different views. Therefore, unsupervised representation learning on multi-view data becomes a compelling topic in machine learning. this paper, we propose novel and flexible model termed Collaborative Multi-View Information Bottleneck Networks (CMIB-Nets), which comprehensively explores the common latent structure view-specific intrinsic information, discards superfluous significantly improving generalization capability of model. Specifically, our proposed relies bottleneck principle to integrate shared among views each view, prompting complete flexibly balancing complementarity consistency multiple We conduct extensive experiments (including analysis, robustness experiment, ablation study) datasets, empirically show promising ability compared state-of-the-arts.