Efficient variational inference for generalized linear mixed models with large datasets

TL;DR

This paper introduces a hybrid Variational Bayes algorithm combined with a divide and recombine strategy to efficiently perform inference on large datasets for generalized linear mixed models, avoiding the need for conjugate priors.

Contribution

It presents a novel hybrid Variational Bayes method with a divide and recombine approach for scalable inference in large datasets, applicable to generalized linear mixed models.

Findings

Demonstrates computational efficiency on simulated datasets

Effective parallel inference for large datasets

Applicable to a wide range of models without conjugacy constraints

Abstract

The article develops a hybrid Variational Bayes algorithm that combines the mean-field and fixed-form Variational Bayes methods. The new estimation algorithm can be used to approximate any posterior without relying on conjugate priors. We propose a divide and recombine strategy for the analysis of large datasets, which partitions a large dataset into smaller pieces and then combines the variational distributions that have been learnt in parallel on each separate piece using the hybrid Variational Bayes algorithm. The proposed method is applied to fitting generalized linear mixed models. The computational efficiency of the parallel and hybrid Variational Bayes algorithm is demonstrated on several simulated and real datasets.