Convex Variational Bayesian Inference for Large Scale Generalized Linear Models
Abstract
We show how variational Bayesian inference can be implemented for very large generalized linear models. Our relaxation is proven to be a convex problem for any log-concave model. We provide a generic double loop algorithm for solving this relaxation on models with arbitrary super-Gaussian potentials. By iteratively decoupling the criterion, most of the work can be done by solving large linear systems, rendering our algorithm orders of magnitude faster than previously proposed solvers for the same problem. We evaluate our method on problems of Bayesian active learning for large binary classification models, and show how to address settings with many candidates and sequential inclusion steps.
Cite
Text
Nickisch and Seeger. "Convex Variational Bayesian Inference for Large Scale Generalized Linear Models." International Conference on Machine Learning, 2009. doi:10.1145/1553374.1553472Markdown
[Nickisch and Seeger. "Convex Variational Bayesian Inference for Large Scale Generalized Linear Models." International Conference on Machine Learning, 2009.](https://mlanthology.org/icml/2009/nickisch2009icml-convex/) doi:10.1145/1553374.1553472BibTeX
@inproceedings{nickisch2009icml-convex,
title = {{Convex Variational Bayesian Inference for Large Scale Generalized Linear Models}},
author = {Nickisch, Hannes and Seeger, Matthias W.},
booktitle = {International Conference on Machine Learning},
year = {2009},
pages = {761-768},
doi = {10.1145/1553374.1553472},
url = {https://mlanthology.org/icml/2009/nickisch2009icml-convex/}
}