A comparison of learning rate selection methods in generalized Bayesian inference

DOI10.1214/21-BA1302arXiv2012.11349OpenAlexW3114174487MaRDI QIDQ6122017FDOQ6122017

Authors:

Publication date: 27 February 2024

Published in: Bayesian Analysis (Search for Journal in Brave)

Abstract: Generalized Bayes posterior distributions are formed by putting a fractional power on the likelihood before combining with the prior via Bayes's formula. This fractional power, which is often viewed as a remedy for potential model misspecification bias, is called the learning rate, and a number of data-driven learning rate selection methods have been proposed in the recent literature. Each of these proposals has a different focus, a different target they aim to achieve, which makes them difficult to compare. In this paper, we provide a direct head-to-head comparison of these learning rate selection methods in various misspecified model scenarios, in terms of several relevant metrics, in particular, coverage probability of the generalized Bayes credible regions. In some examples all the methods perform well, while in others the misspecification is too severe to be overcome, but we find that the so-called generalized posterior calibration algorithm tends to outperform the others in terms of credible region coverage probability.

Full work available at URL: https://arxiv.org/abs/2012.11349

zbMATH Keywords

model misspecificationcoverage probabilitygeneralized posterior calibration algorithmSafeBayes algorithm

Mathematics Subject Classification ID

Statistics (62-XX) Interacting random processes; statistical mechanics type models; percolation theory (60K35)

Cites Work

• Empirical Bayes posterior concentration in sparse high-dimensional linear models
• Title not available (Why is that?)
• Bayesian Inference for Logistic Models Using Pólya–Gamma Latent Variables
• Title not available (Why is that?)
• Bayesian fractional posteriors
• Asymptotic Statistics
• Fundamentals of Nonparametric Bayesian Inference
• On the consistency of Bayes estimates
• On inconsistent Bayes estimates of location
• Convergence rates of posterior distributions.
• Misspecification in infinite-dimensional Bayesian statistics
• Title not available (Why is that?)
• Limiting Behavior of Posterior Distributions when the Model is Incorrect
• Inconsistency of Bayesian inference for misspecified linear models, and a proposal for repairing it
• Robust Bayesian Inference via Coarsening
• A General Framework for Updating Belief Distributions
• Asymptotic behavior of Bayes estimates under possibly incorrect models
• Is Bayes posterior just quick and dirty confidence?
• Posterior consistency of Dirichlet mixtures in density estimation
• The Bernstein-von Mises theorem under misspecification
• The consistency of posterior distributions in nonparametric problems
• On posterior concentration in misspecified models
• On Bayesian consistency
• Data tracking and the understanding of Bayesian consistency
• Title not available (Why is that?)
• Bayesian inference with misspecified models
• From \(\varepsilon\)-entropy to KL-entropy: analysis of minimum information complexity density estima\-tion
• Safe probability
• The safe Bayesian. Learning the learning rate via the mixability gap
• OUP accepted manuscript
• Gibbs posterior inference on the minimum clinically important difference
• Title not available (Why is that?)
• Gibbs posterior inference on multivariate quantiles
• Model-free posterior inference on the area under the receiver operating characteristic curve
• Robust and rate-optimal Gibbs posterior inference on the boundary of a noisy image
• Empirical priors and coverage of posterior credible sets in a sparse normal mean model
• False confidence, non-additive beliefs, and valid statistical inference
• Calibrating general posterior credible regions
• General Bayesian updating and the loss-likelihood bootstrap
• Data-driven priors and their posterior concentration rates
• Comment: ``Bayes, oracle Bayes and empirical Bayes
• Title not available (Why is that?)
• Minimum clinically important difference in medical studies

Cited In (2)

• A reduced-rank approach to predicting multiple binary responses through machine learning
• Sampling algorithms in statistical physics: a guide for statistics and machine learning

This page was built for publication: A comparison of learning rate selection methods in generalized Bayesian inference

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6122017)