Improving Variational Autoencoder Estimation from Incomplete Data with Mixture Variational Families
Abstract
We consider the task of estimating variational autoencoders (VAEs) when the training data is incomplete. We show that missing data increases the complexity of the model’s posterior distribution over the latent variables compared to the fully-observed case. The increased complexity may adversely affect the fit of the model due to a mismatch between the variational and model posterior distributions. We introduce two strategies based on (i) finite variational-mixture and (ii) imputation-based variational-mixture distributions to address the increased posterior complexity. Through a comprehensive evaluation of the proposed approaches, we show that variational mixtures are effective at improving the accuracy of VAE estimation from incomplete data.
Cite
Text
Simkus and Gutmann. "Improving Variational Autoencoder Estimation from Incomplete Data with Mixture Variational Families." Transactions on Machine Learning Research, 2024.Markdown
[Simkus and Gutmann. "Improving Variational Autoencoder Estimation from Incomplete Data with Mixture Variational Families." Transactions on Machine Learning Research, 2024.](https://mlanthology.org/tmlr/2024/simkus2024tmlr-improving/)BibTeX
@article{simkus2024tmlr-improving,
title = {{Improving Variational Autoencoder Estimation from Incomplete Data with Mixture Variational Families}},
author = {Simkus, Vaidotas and Gutmann, Michael U.},
journal = {Transactions on Machine Learning Research},
year = {2024},
url = {https://mlanthology.org/tmlr/2024/simkus2024tmlr-improving/}
}