Green Generative Modeling: Recycling Dirty Data using Recurrent Variational Autoencoders

Yu Wang, Bin Dai, Gang Hua, John Aston, David Wipf
Proceedings of the 33rd Conference on Uncertainty in Artificial Intelligence, PMLR R15:391-400, 2017.

Abstract

This paper explores two useful modifications of the recent variational autoencoder (VAE), a popular deep generative modeling framework that dresses traditional autoencoders with probabilistic attire. The first involves a specially-tailored form of conditioning that allows us to simplify the VAE decoder structure while simultaneously introducing robustness to outliers. In a related vein, a second, complementary alteration is proposed to further build invariance to contaminated or dirty samples via a data augmentation process that amounts to recycling. In brief, to the extent that the VAE is legitimately a representative generative model, then each output from the decoder should closely resemble an authentic sample, which can then be resubmitted as a novel input ad infinitum. Moreover, this can be accomplished via special recurrent connections without the need for additional parameters to be trained. We evaluate these proposals on multiple practical outlier-removal and generative modeling tasks, demonstrating considerable improvements over existing algorithms.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR15-wang17a, title = {Green Generative Modeling: Recycling Dirty Data using Recurrent Variational Autoencoders}, author = {Wang, Yu and Dai, Bin and Hua, Gang and Aston, John and Wipf, David}, booktitle = {Proceedings of the 33rd Conference on Uncertainty in Artificial Intelligence}, pages = {391--400}, year = {2017}, editor = {Elidan, Gal and Kersting, Kristian}, volume = {R15}, series = {Proceedings of Machine Learning Research}, month = {11--15 Aug}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r15/main/assets/wang17a/wang17a.pdf}, url = {https://proceedings.mlr.press/r15/wang17a.html}, abstract = {This paper explores two useful modifications of the recent variational autoencoder (VAE), a popular deep generative modeling framework that dresses traditional autoencoders with probabilistic attire. The first involves a specially-tailored form of conditioning that allows us to simplify the VAE decoder structure while simultaneously introducing robustness to outliers. In a related vein, a second, complementary alteration is proposed to further build invariance to contaminated or dirty samples via a data augmentation process that amounts to recycling. In brief, to the extent that the VAE is legitimately a representative generative model, then each output from the decoder should closely resemble an authentic sample, which can then be resubmitted as a novel input ad infinitum. Moreover, this can be accomplished via special recurrent connections without the need for additional parameters to be trained. We evaluate these proposals on multiple practical outlier-removal and generative modeling tasks, demonstrating considerable improvements over existing algorithms.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Green Generative Modeling: Recycling Dirty Data using Recurrent Variational Autoencoders %A Yu Wang %A Bin Dai %A Gang Hua %A John Aston %A David Wipf %B Proceedings of the 33rd Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2017 %E Gal Elidan %E Kristian Kersting %F pmlr-vR15-wang17a %I PMLR %P 391--400 %U https://proceedings.mlr.press/r15/wang17a.html %V R15 %X This paper explores two useful modifications of the recent variational autoencoder (VAE), a popular deep generative modeling framework that dresses traditional autoencoders with probabilistic attire. The first involves a specially-tailored form of conditioning that allows us to simplify the VAE decoder structure while simultaneously introducing robustness to outliers. In a related vein, a second, complementary alteration is proposed to further build invariance to contaminated or dirty samples via a data augmentation process that amounts to recycling. In brief, to the extent that the VAE is legitimately a representative generative model, then each output from the decoder should closely resemble an authentic sample, which can then be resubmitted as a novel input ad infinitum. Moreover, this can be accomplished via special recurrent connections without the need for additional parameters to be trained. We evaluate these proposals on multiple practical outlier-removal and generative modeling tasks, demonstrating considerable improvements over existing algorithms. %Z Reissued by PMLR on 04 October 2026.
APA
Wang, Y., Dai, B., Hua, G., Aston, J. & Wipf, D.. (2017). Green Generative Modeling: Recycling Dirty Data using Recurrent Variational Autoencoders. Proceedings of the 33rd Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R15:391-400 Available from https://proceedings.mlr.press/r15/wang17a.html. Reissued by PMLR on 04 October 2026.

Related Material