Deep Incomplete Multi-view Learning via Cyclic Permutation of VAEs
This addresses the challenge of irregularly missing views in multi-view learning, which is incremental as it builds on existing VAE-based methods to improve representation sufficiency and consistency.
The paper tackles the problem of incomplete multi-view data leading to insufficient and inconsistent representations by proposing MVP, a method using cyclic permutations of VAEs to infer missing views and aggregate information, achieving superior performance in clustering and generation tasks across seven datasets with varying missing ratios.
Multi-View Representation Learning (MVRL) aims to derive a unified representation from multi-view data by leveraging shared and complementary information across views. However, when views are irregularly missing, the incomplete data can lead to representations that lack sufficiency and consistency. To address this, we propose Multi-View Permutation of Variational Auto-Encoders (MVP), which excavates invariant relationships between views in incomplete data. MVP establishes inter-view correspondences in the latent space of Variational Auto-Encoders, enabling the inference of missing views and the aggregation of more sufficient information. To derive a valid Evidence Lower Bound (ELBO) for learning, we apply permutations to randomly reorder variables for cross-view generation and then partition them by views to maintain invariant meanings under permutations. Additionally, we enhance consistency by introducing an informational prior with cyclic permutations of posteriors, which turns the regularization term into a similarity measure across distributions. We demonstrate the effectiveness of our approach on seven diverse datasets with varying missing ratios, achieving superior performance in multi-view clustering and generation tasks.