Preventing posterior collapse in variational autoencoders for text generation via decoder regularization
Year: 2021
Type: preprint
Abstract: Variational autoencoders trained to minimize the reconstruction error are sensitive to the posterior collapse problem, that is the proposal posterior distribution is always equal to the prior. We propose a novel regularization method based on fraternal dropout to prevent posterior collapse. We evaluate our approach using several metrics and observe improvements in all the tested configurations.
Source: arXiv (Cornell University)
Authors: Petit, Alban, Corro, Caio
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3211080416