Variational Boosting: Iteratively Refining Posterior Approximations

Abstract

We propose a black-box variational inference method to approximate intractable distributions with an increasingly rich approximating class. Our method, variational boosting, iteratively refines an existing variational approximation by solving a sequence of optimization problems, allowing a trade-off between computation time and accuracy. We expand the variational approximating class by incorporating additional covariance structure and by introducing new components to form a mixture. We apply variational boosting to synthetic and real statistical models, and show that the resulting posterior inferences compare favorably to existing variational algorithms.

Cite

Text

Miller et al. "Variational Boosting: Iteratively Refining Posterior Approximations." International Conference on Machine Learning, 2017.

Markdown

[Miller et al. "Variational Boosting: Iteratively Refining Posterior Approximations." International Conference on Machine Learning, 2017.](https://mlanthology.org/icml/2017/miller2017icml-variational/)

BibTeX

@inproceedings{miller2017icml-variational,
  title     = {{Variational Boosting: Iteratively Refining Posterior Approximations}},
  author    = {Miller, Andrew C. and Foti, Nicholas J. and Adams, Ryan P.},
  booktitle = {International Conference on Machine Learning},
  year      = {2017},
  pages     = {2420-2429},
  volume    = {70},
  url       = {https://mlanthology.org/icml/2017/miller2017icml-variational/}
}