I presented a paper at ICPR 2024 in Kolkata, India. The paper distinguishes between the often misused important parameters and in variational autoencoders.
Title: Beta-Sigma VAE: Separating beta and decoder variance in Gaussian variational autoencoder
Authors: Seunghwan Kim, Seungkyu Lee
Abstract: Variational autoencoder (VAE) is an established generative model but is notorious for its blurriness. In this work, we investigate the blurry output problem of VAE and resolve it, exploiting the variance of Gaussian decoder and of beta-VAE [Higgins et al. 2016]. Specifically, we reveal that the indistinguishability of decoder variance and hinders appropriate analysis of the model by random likelihood value, and limits performance improvement by omitting the gain from . To address the problem, we propose Beta-Sigma VAE (BS-VAE) that explicitly separates and decoder variance in the model. Our method demonstrates not only superior performance in natural image synthesis but also controllable parameters and predictable analysis compared to conventional VAE. In our experimental evaluation, we employ the analysis of rate-distortion curve and proxy metrics on computer vision datasets. The code is available on https://github.com/overnap/BS-VAE
