Meta-Learning Divergences for Variational Inference

Ruqi Zhang, Yingzhen Li, Christopher De Sa, Sam Devlin, Cheng Zhang
Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, PMLR 130:4024-4032, 2021.

Abstract

Variational inference (VI) plays an essential role in approximate Bayesian inference due to its computational efficiency and broad applicability. Crucial to the performance of VI is the selection of the associated divergence measure, as VI approximates the intractable distribution by minimizing this divergence. In this paper we propose a meta-learning algorithm to learn the divergence metric suited for the task of interest, automating the design of VI methods. In addition, we learn the initialization of the variational parameters without additional cost when our method is deployed in the few-shot learning scenarios. We demonstrate our approach outperforms standard VI on Gaussian mixture distribution approximation, Bayesian neural network regression, image generation with variational autoencoders and recommender systems with a partial variational autoencoder.

Cite this Paper


BibTeX
@InProceedings{pmlr-v130-zhang21o, title = { Meta-Learning Divergences for Variational Inference }, author = {Zhang, Ruqi and Li, Yingzhen and De Sa, Christopher and Devlin, Sam and Zhang, Cheng}, booktitle = {Proceedings of The 24th International Conference on Artificial Intelligence and Statistics}, pages = {4024--4032}, 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/zhang21o/zhang21o.pdf}, url = {https://proceedings.mlr.press/v130/zhang21o.html}, abstract = { Variational inference (VI) plays an essential role in approximate Bayesian inference due to its computational efficiency and broad applicability. Crucial to the performance of VI is the selection of the associated divergence measure, as VI approximates the intractable distribution by minimizing this divergence. In this paper we propose a meta-learning algorithm to learn the divergence metric suited for the task of interest, automating the design of VI methods. In addition, we learn the initialization of the variational parameters without additional cost when our method is deployed in the few-shot learning scenarios. We demonstrate our approach outperforms standard VI on Gaussian mixture distribution approximation, Bayesian neural network regression, image generation with variational autoencoders and recommender systems with a partial variational autoencoder. } }
Endnote
%0 Conference Paper %T Meta-Learning Divergences for Variational Inference %A Ruqi Zhang %A Yingzhen Li %A Christopher De Sa %A Sam Devlin %A Cheng Zhang %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-zhang21o %I PMLR %P 4024--4032 %U https://proceedings.mlr.press/v130/zhang21o.html %V 130 %X Variational inference (VI) plays an essential role in approximate Bayesian inference due to its computational efficiency and broad applicability. Crucial to the performance of VI is the selection of the associated divergence measure, as VI approximates the intractable distribution by minimizing this divergence. In this paper we propose a meta-learning algorithm to learn the divergence metric suited for the task of interest, automating the design of VI methods. In addition, we learn the initialization of the variational parameters without additional cost when our method is deployed in the few-shot learning scenarios. We demonstrate our approach outperforms standard VI on Gaussian mixture distribution approximation, Bayesian neural network regression, image generation with variational autoencoders and recommender systems with a partial variational autoencoder.
APA
Zhang, R., Li, Y., De Sa, C., Devlin, S. & Zhang, C.. (2021). Meta-Learning Divergences for Variational Inference . Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 130:4024-4032 Available from https://proceedings.mlr.press/v130/zhang21o.html.

Related Material