Posterior concentration and fast convergence rates for generalized Bayesian learning
From MaRDI portal
Publication:2666765
Abstract: In this paper, we study the learning rate of generalized Bayes estimators in a general setting where the hypothesis class can be uncountable and have an irregular shape, the loss function can have heavy tails, and the optimal hypothesis may not be unique. We prove that under the multi-scale Bernstein's condition, the generalized posterior distribution concentrates around the set of optimal hypotheses and the generalized Bayes estimator can achieve fast learning rate. Our results are applied to show that the standard Bayesian linear regression is robust to heavy-tailed distributions.
Recommendations
- Fast rates for general unbounded loss functions: from ERM to generalized Bayes
- The safe Bayesian. Learning the learning rate via the mixability gap
- Convergence rates of posterior distributions.
- Convergence rates for posterior distributions and adaptive estimation
- Fast rates in statistical and online learning
Cites work
- Asymptotic behaviour of the posterior distribution in overfitted mixture models
- Bayesian Inference for Hospital Quality in a Selection Model
- Bayesian optimal adaptive estimation using a sieve prior
- Birth/birth-death processes and their computable transition probabilities with biological applications
- Borrowing strengh in hierarchical Bayes: posterior concentration of the Dirichlet base measure
- Convergence of Griddy Gibbs sampling and other perturbed Markov chains
- Convergence rates of least squares regression estimators with heavy-tailed errors
- Convergence rates of posterior distributions.
- Direct likelihood-based inference for discretely observed stochastic compartmental models of infectious disease
- Empirical minimization
- Empirical risk minimization for heavy-tailed losses
- Fast learning rates for plug-in classifiers
- Fast rates in statistical and online learning
- From -entropy to KL-entropy: analysis of minimum information complexity density estima\-tion
- General nonexact oracle inequalities for classes with a subexponential envelope
- Generalization and Robustness of Batched Weighted Average Algorithm with V-Geometrically Ergodic Markov Data
- Generalization bounds for averaged classifiers
- Inconsistency of Bayesian inference for misspecified linear models, and a proposal for repairing it
- Information-theoretic upper and lower bounds for statistical estimation
- Learning from non-iid data: fast rates for the one-vs-all multiclass plug-in classifiers
- Learning without concentration
- Loss minimization and parameter estimation with heavy tails
- Mean estimation and regression under heavy-tailed distributions: A survey
- On consistency and robustness properties of support vector machines for heavy-tailed distributions
- The safe Bayesian. Learning the learning rate via the mixability gap
Cited in
(5)- The safe Bayesian. Learning the learning rate via the mixability gap
- Fast rates for general unbounded loss functions: from ERM to generalized Bayes
- Asymptotic normality, concentration, and coverage of generalized posteriors
- Generalized Bayes Quantification Learning under Dataset Shift
- On generalized computable universal priors and their convergence
This page was built for publication: Posterior concentration and fast convergence rates for generalized Bayesian learning
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2666765)
