Controlling Posterior Collapse by an Inverse Lipschitz Constraint on the Decoder Network
Abstract
Variational autoencoders (VAEs) are one of the deep generative models that have experienced enormous success over the past decades. However, in practice, they suffer from a problem called posterior collapse, which occurs when the posterior distribution coincides, or collapses, with the prior taking no information from the latent structure of the input data into consideration. In this work, we introduce an inverse Lipschitz neural network into the decoder and, based on this architecture, provide a new method that can control in a simple and clear manner the degree of posterior collapse for a wide range of VAE models equipped with a concrete theoretical guarantee. We also illustrate the effectiveness of our method through several numerical experiments.
Cite
Text
Kinoshita et al. "Controlling Posterior Collapse by an Inverse Lipschitz Constraint on the Decoder Network." International Conference on Machine Learning, 2023.Markdown
[Kinoshita et al. "Controlling Posterior Collapse by an Inverse Lipschitz Constraint on the Decoder Network." International Conference on Machine Learning, 2023.](https://mlanthology.org/icml/2023/kinoshita2023icml-controlling/)BibTeX
@inproceedings{kinoshita2023icml-controlling,
title = {{Controlling Posterior Collapse by an Inverse Lipschitz Constraint on the Decoder Network}},
author = {Kinoshita, Yuri and Oono, Kenta and Fukumizu, Kenji and Yoshida, Yuichi and Maeda, Shin-Ichi},
booktitle = {International Conference on Machine Learning},
year = {2023},
pages = {17041-17060},
volume = {202},
url = {https://mlanthology.org/icml/2023/kinoshita2023icml-controlling/}
}