2000 character limit reached
Using VAEs to Learn Latent Variables: Observations on Applications in cryo-EM (2303.07487v2)
Published 13 Mar 2023 in stat.ML, cs.LG, and q-bio.QM
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.