Using VAEs to Learn Latent Variables: Observations on Applications in cryo-EM

TL;DR

This paper examines how variational autoencoders (VAEs) can learn latent variables in biological applications, revealing that the encoder's representations resemble traditional explicit latent variable models.

Contribution

It provides a qualitative analysis of VAE encoder properties in cryo-EM, highlighting similarities to explicit latent representations in biological data modeling.

Findings

Encoder resembles traditional explicit latent variables

VAEs effectively characterize biological systems

Qualitative insights into VAE amortization in cryo-EM

Abstract

Variational autoencoders (VAEs) are a popular generative model used to approximate distributions. The encoder part of the VAE is used in amortized learning of latent variables, producing a latent representation for data samples. Recently, VAEs have been used to characterize physical and biological systems. In this case study, we qualitatively examine the amortization properties of a VAE used in biological applications. We find that in this application the encoder bears a qualitative resemblance to more traditional explicit representation of latent variables.