SIReN-VAE: Leveraging Flows and Amortized Inference for Bayesian Networks

TL;DR

This paper introduces SIReN-VAE, a novel variational autoencoder that integrates Bayesian network structures with residual flows to model complex dependencies in latent variables, enhancing flexibility and performance.

Contribution

It extends VAEs by incorporating Bayesian network-based dependency structures into both prior and inference networks using graphical residual flows, enabling richer latent representations.

Findings

Improved modeling of complex dependencies in synthetic datasets

Enhanced performance in data-sparse scenarios

Effective encoding of conditional independence through residual flows

Abstract

Initial work on variational autoencoders assumed independent latent variables with simple distributions. Subsequent work has explored incorporating more complex distributions and dependency structures: including normalizing flows in the encoder network allows latent variables to entangle non-linearly, creating a richer class of distributions for the approximate posterior, and stacking layers of latent variables allows more complex priors to be specified for the generative model. This work explores incorporating arbitrary dependency structures, as specified by Bayesian networks, into VAEs. This is achieved by extending both the prior and inference network with graphical residual flows - residual flows that encode conditional independence by masking the weight matrices of the flow's residual blocks. We compare our model's performance on several synthetic datasets and show its potential in data-sparse settings.