A Semi-supervised Generative Model for Incomplete Multi-view Data Integration with Missing Labels

TL;DR

This paper introduces a semi-supervised generative model that effectively integrates incomplete multi-view data with missing labels, improving prediction and imputation by leveraging both labeled and unlabeled samples.

Contribution

It proposes a novel semi-supervised probabilistic framework that combines information bottleneck and mutual information maximization for multi-view data with missing views and labels.

Findings

Outperforms existing methods in multi-omics data imputation.

Achieves higher predictive accuracy with limited labeled data.

Effectively leverages unlabeled data for better representation learning.

Abstract

Multi-view learning is widely applied to real-life datasets, such as multiple omics biological data, but it often suffers from both missing views and missing labels. Prior probabilistic approaches addressed the missing view problem by using a product-of-experts scheme to aggregate representations from present views and achieved superior performance over deterministic classifiers, using the information bottleneck (IB) principle. However, the IB framework is inherently fully supervised and cannot leverage unlabeled data. In this work, we propose a semi-supervised generative model that utilizes both labeled and unlabeled samples in a unified framework. Our method maximizes the likelihood of unlabeled samples to learn a latent space shared with the IB on labeled data. We also perform cross-view mutual information maximization in the latent space to enhance the extraction of shared information across views. Compared to existing approaches, our model achieves better predictive and imputation performance on both image and multi-omics data with missing views and limited labeled samples.