Scalable Gaussian Process Variational Autoencoders

Metod Jazbec, Matt Ashman, Vincent Fortuin, Michael Pearce, Stephan Mandt, Gunnar Rätsch
Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, PMLR 130:3511-3519, 2021.

Abstract

Conventional variational autoencoders fail in modeling correlations between data points due to their use of factorized priors. Amortized Gaussian process inference through GP-VAEs has led to significant improvements in this regard, but is still inhibited by the intrinsic complexity of exact GP inference. We improve the scalability of these methods through principled sparse inference approaches. We propose a new scalable GP-VAE model that outperforms existing approaches in terms of runtime and memory footprint, is easy to implement, and allows for joint end-to-end optimization of all components.

Cite this Paper

BibTeX
@InProceedings{pmlr-v130-jazbec21a, title = { Scalable Gaussian Process Variational Autoencoders }, author = {Jazbec, Metod and Ashman, Matt and Fortuin, Vincent and Pearce, Michael and Mandt, Stephan and R{\"a}tsch, Gunnar}, booktitle = {Proceedings of The 24th International Conference on Artificial Intelligence and Statistics}, pages = {3511--3519}, year = {2021}, editor = {Banerjee, Arindam and Fukumizu, Kenji}, volume = {130}, series = {Proceedings of Machine Learning Research}, month = {13--15 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v130/jazbec21a/jazbec21a.pdf}, url = {https://proceedings.mlr.press/v130/jazbec21a.html}, abstract = { Conventional variational autoencoders fail in modeling correlations between data points due to their use of factorized priors. Amortized Gaussian process inference through GP-VAEs has led to significant improvements in this regard, but is still inhibited by the intrinsic complexity of exact GP inference. We improve the scalability of these methods through principled sparse inference approaches. We propose a new scalable GP-VAE model that outperforms existing approaches in terms of runtime and memory footprint, is easy to implement, and allows for joint end-to-end optimization of all components. } }
Endnote
%0 Conference Paper %T Scalable Gaussian Process Variational Autoencoders %A Metod Jazbec %A Matt Ashman %A Vincent Fortuin %A Michael Pearce %A Stephan Mandt %A Gunnar Rätsch %B Proceedings of The 24th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2021 %E Arindam Banerjee %E Kenji Fukumizu %F pmlr-v130-jazbec21a %I PMLR %P 3511--3519 %U https://proceedings.mlr.press/v130/jazbec21a.html %V 130 %X Conventional variational autoencoders fail in modeling correlations between data points due to their use of factorized priors. Amortized Gaussian process inference through GP-VAEs has led to significant improvements in this regard, but is still inhibited by the intrinsic complexity of exact GP inference. We improve the scalability of these methods through principled sparse inference approaches. We propose a new scalable GP-VAE model that outperforms existing approaches in terms of runtime and memory footprint, is easy to implement, and allows for joint end-to-end optimization of all components.
APA
Jazbec, M., Ashman, M., Fortuin, V., Pearce, M., Mandt, S. & Rätsch, G.. (2021). Scalable Gaussian Process Variational Autoencoders . Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 130:3511-3519 Available from https://proceedings.mlr.press/v130/jazbec21a.html.

Related Material