Implementing and Automating Fixed-Form Variational Posterior Approximation through Stochastic Linear Regression

2014/01/09 by Tim Salimans, Salimans, Tim · 1 citation
Computer Science · Decision Sciences · #Advanced Multi-Objective Optimization Algorithms #Computation (stat.CO) #FOS: Computer and information sciences #Gaussian Processes and Bayesian Inference #Probabilistic and Robust Engineering Design

paper · pdf · doi:10.48550/arxiv.1401.2135

openalex publication_date 2014/01/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We recently proposed a general algorithm for approximating nonstandard Bayesian posterior distributions by minimization of their Kullback-Leibler divergence with respect to a more convenient approximating distribution. In this note we offer details on how to efficiently implement this algorithm in practice. We also suggest default choices for the form of the posterior approximation, the number of iterations, the step size, and other user choices. By using these defaults it becomes possible to construct good posterior approximations for hierarchical models completely automatically.

Cited by

• Rao-Blackwellised Reparameterisation Gradients

Related

• On Using Control Variates with Stochastic Approximation for Variational Bayes and its Connection to Stochastic Linear Regression
• Variational approximations using Fisher divergence
• Skew-symmetric approximations of posterior distributions
• On the Robustness to Misspecification of α-Posteriors and Their Variational Approximations
• Understanding Variational Inference in Function-Space