Controlling the Interaction Between Generation and Inference in Semi-Supervised Variational Autoencoders Using Importance Weighting

TL;DR

This paper analyzes semi-supervised VAEs, revealing how they use the posterior to guide inference, and introduces importance weighting techniques to better control the influence of observed and unobserved latent variables, improving performance.

Contribution

It provides a new understanding of semi-supervised VAEs and proposes importance weighting objectives for finer control over latent variable influence.

Findings

Improved classification accuracy on IMDB and AG News datasets.

Finer control over the influence of latent variables enhances training stability.

Behavior aligns with theoretical analysis of semi-supervised VAE mechanisms.

Abstract

Even though Variational Autoencoders (VAEs) are widely used for semi-supervised learning, the reason why they work remains unclear. In fact, the addition of the unsupervised objective is most often vaguely described as a regularization. The strength of this regularization is controlled by down-weighting the objective on the unlabeled part of the training set. Through an analysis of the objective of semi-supervised VAEs, we observe that they use the posterior of the learned generative model to guide the inference model in learning the partially observed latent variable. We show that given this observation, it is possible to gain finer control on the effect of the unsupervised objective on the training procedure. Using importance weighting, we derive two novel objectives that prioritize either one of the partially observed latent variable, or the unobserved latent variable. Experiments on the IMDB english sentiment analysis dataset and on the AG News topic classification dataset show the improvements brought by our prioritization mechanism and exhibit a behavior that is inline with our description of the inner working of Semi-Supervised VAEs.