Multiple Importance Sampling ELBO and Deep Ensembles of Variational Approximations

Oskar Kviman, Harald Melin, Hazal Koptagel, Victor Elvira, Jens Lagergren
Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, PMLR 151:10687-10702, 2022.

Abstract

In variational inference (VI), the marginal log-likelihood is estimated using the standard evidence lower bound (ELBO), or improved versions as the importance weighted ELBO (IWELBO). We propose the multiple importance sampling ELBO (MISELBO), a versatile yet simple framework. MISELBO is applicable in both amortized and classical VI, and it uses ensembles, e.g., deep ensembles, of independently inferred variational approximations. As far as we are aware, the concept of deep ensembles in amortized VI has not previously been established. We prove that MISELBO provides a tighter bound than the average of standard ELBOs, and demonstrate empirically that it gives tighter bounds than the average of IWELBOs. MISELBO is evaluated in density-estimation experiments that include MNIST and several real-data phylogenetic tree inference problems. First, on the MNIST dataset, MISELBO boosts the density-estimation performances of a state-of-the-art model, nouveau VAE. Second, in the phylogenetic tree inference setting, our framework enhances a state-of-the-art VI algorithm that uses normalizing flows. On top of the technical benefits of MISELBO, it allows to unveil connections between VI and recent advances in the importance sampling literature, paving the way for further methodological advances. We provide our code at https://github.com/Lagergren-Lab/MISELBO.

Cite this Paper


BibTeX
@InProceedings{pmlr-v151-kviman22a, title = { Multiple Importance Sampling ELBO and Deep Ensembles of Variational Approximations }, author = {Kviman, Oskar and Melin, Harald and Koptagel, Hazal and Elvira, Victor and Lagergren, Jens}, booktitle = {Proceedings of The 25th International Conference on Artificial Intelligence and Statistics}, pages = {10687--10702}, year = {2022}, editor = {Camps-Valls, Gustau and Ruiz, Francisco J. R. and Valera, Isabel}, volume = {151}, series = {Proceedings of Machine Learning Research}, month = {28--30 Mar}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v151/kviman22a/kviman22a.pdf}, url = {https://proceedings.mlr.press/v151/kviman22a.html}, abstract = { In variational inference (VI), the marginal log-likelihood is estimated using the standard evidence lower bound (ELBO), or improved versions as the importance weighted ELBO (IWELBO). We propose the multiple importance sampling ELBO (MISELBO), a versatile yet simple framework. MISELBO is applicable in both amortized and classical VI, and it uses ensembles, e.g., deep ensembles, of independently inferred variational approximations. As far as we are aware, the concept of deep ensembles in amortized VI has not previously been established. We prove that MISELBO provides a tighter bound than the average of standard ELBOs, and demonstrate empirically that it gives tighter bounds than the average of IWELBOs. MISELBO is evaluated in density-estimation experiments that include MNIST and several real-data phylogenetic tree inference problems. First, on the MNIST dataset, MISELBO boosts the density-estimation performances of a state-of-the-art model, nouveau VAE. Second, in the phylogenetic tree inference setting, our framework enhances a state-of-the-art VI algorithm that uses normalizing flows. On top of the technical benefits of MISELBO, it allows to unveil connections between VI and recent advances in the importance sampling literature, paving the way for further methodological advances. We provide our code at https://github.com/Lagergren-Lab/MISELBO. } }
Endnote
%0 Conference Paper %T Multiple Importance Sampling ELBO and Deep Ensembles of Variational Approximations %A Oskar Kviman %A Harald Melin %A Hazal Koptagel %A Victor Elvira %A Jens Lagergren %B Proceedings of The 25th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2022 %E Gustau Camps-Valls %E Francisco J. R. Ruiz %E Isabel Valera %F pmlr-v151-kviman22a %I PMLR %P 10687--10702 %U https://proceedings.mlr.press/v151/kviman22a.html %V 151 %X In variational inference (VI), the marginal log-likelihood is estimated using the standard evidence lower bound (ELBO), or improved versions as the importance weighted ELBO (IWELBO). We propose the multiple importance sampling ELBO (MISELBO), a versatile yet simple framework. MISELBO is applicable in both amortized and classical VI, and it uses ensembles, e.g., deep ensembles, of independently inferred variational approximations. As far as we are aware, the concept of deep ensembles in amortized VI has not previously been established. We prove that MISELBO provides a tighter bound than the average of standard ELBOs, and demonstrate empirically that it gives tighter bounds than the average of IWELBOs. MISELBO is evaluated in density-estimation experiments that include MNIST and several real-data phylogenetic tree inference problems. First, on the MNIST dataset, MISELBO boosts the density-estimation performances of a state-of-the-art model, nouveau VAE. Second, in the phylogenetic tree inference setting, our framework enhances a state-of-the-art VI algorithm that uses normalizing flows. On top of the technical benefits of MISELBO, it allows to unveil connections between VI and recent advances in the importance sampling literature, paving the way for further methodological advances. We provide our code at https://github.com/Lagergren-Lab/MISELBO.
APA
Kviman, O., Melin, H., Koptagel, H., Elvira, V. & Lagergren, J.. (2022). Multiple Importance Sampling ELBO and Deep Ensembles of Variational Approximations . Proceedings of The 25th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 151:10687-10702 Available from https://proceedings.mlr.press/v151/kviman22a.html.

Related Material