On the properties of variational approximations of Gibbs posteriors

zbMath1437.62129arXiv1506.04091MaRDI QIDQ2958606

James Ridgway, Pierre Alquier, Nicolas Chopin

Publication date: 3 February 2017

Full work available at URL: https://arxiv.org/abs/1506.04091

zbMATH Keywords

Markov chain Monte Carlobig datasetsnon-asymptotic risk boundPAC-Bayesian approachrandom estimatorvariational approximations of the Gibbs posterior

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08) Density estimation (62G07) Bayesian problems; characterization of Bayes procedures (62C10) Statistical aspects of big data and data science (62R07)

Related Items (23)

1-bit matrix completion: PAC-Bayesian analysis of a variational approximation ⋮ Simpler PAC-Bayesian bounds for hostile data ⋮ Unnamed Item ⋮ An oracle inequality for quasi-Bayesian nonnegative matrix factorization ⋮ Model-free posterior inference on the area under the receiver operating characteristic curve ⋮ A reduced-rank approach to predicting multiple binary responses through machine learning ⋮ Explicit construction of the minimum error variance estimator for stochastic LTI-ss systems ⋮ User-friendly Introduction to PAC-Bayes Bounds ⋮ A global-local approach for detecting hotspots in multiple-response regression ⋮ \(\alpha\)-variational inference with statistical guarantees ⋮ Concentration of tempered posteriors and of their variational approximations ⋮ An efficient adaptive MCMC algorithm for pseudo-Bayesian quantum tomography ⋮ Variational inference for cutting feedback in misspecified models ⋮ Adaptive variational Bayes: optimality, computation and applications ⋮ Gibbs posterior concentration rates under sub-exponential type losses ⋮ Bayesian fractional posteriors ⋮ Unnamed Item ⋮ Pseudo-Bayesian quantum tomography with rank-adaptation ⋮ Stability of Gibbs Posteriors from the Wasserstein Loss for Bayesian Full Waveform Inversion ⋮ Consistency of variational Bayes inference for estimation and model selection in mixtures ⋮ Leave Pima Indians alone: binary regression as a benchmark for Bayesian computation ⋮ Estimation bounds and sharp oracle inequalities of regularized procedures with Lipschitz loss functions ⋮ Frequentist Consistency of Variational Bayes

Uses Software

• mlbench
• PRMLT

Cites Work

• Unnamed Item
• Nuclear-norm penalization and optimal rates for noisy low-rank matrix completion
• Robust linear least squares regression
• Gibbs posterior for variable selection in high-dimensional classification and data mining
• Statistical behavior and consistency of classification methods based on convex risk minimization.
• An introduction to variational methods for graphical models
• A Bayesian approach for noisy matrix completion: optimal rate under general sampling distribution
• Bayesian Methods for Low-Rank Matrix Estimation: Short Survey and Theoretical Study
• Information-theoretic upper and lower bounds for statistical estimation
• Sparse single-index model
• A Class of Statistics with Asymptotically Normal Distribution

This page was built for publication: On the properties of variational approximations of Gibbs posteriors