Variational Interpretable Learning from Multi-view Data

TL;DR

This paper introduces DICCA, a deep variational model for multi-view data that disentangles shared and view-specific variations, enhancing interpretability and performance over existing methods.

Contribution

It extends linear CCA to a nonlinear, deep generative framework with structured sparsity for interpretability in multi-view learning.

Findings

Competitive performance on real-world datasets

Effective disentanglement of shared and view-specific factors

Enhanced interpretability through sparsity priors

Abstract

The main idea of canonical correlation analysis (CCA) is to map different views onto a common latent space with maximum correlation. We propose a deep interpretable variational canonical correlation analysis (DICCA) for multi-view learning. The developed model extends the existing latent variable model for linear CCA to nonlinear models through the use of deep generative networks. DICCA is designed to disentangle both the shared and view-specific variations for multi-view data. To further make the model more interpretable, we place a sparsity-inducing prior on the latent weight with a structured variational autoencoder that is comprised of view-specific generators. Empirical results on real-world datasets show that our methods are competitive across domains.