Variance reduction for dependent sequences with applications to stochastic gradient MCMC
From MaRDI portal
Abstract: In this paper we propose a novel and practical variance reduction approach for additive functionals of dependent sequences. Our approach combines the use of control variates with the minimisation of an empirical variance estimate. We analyse finite sample properties of the proposed method and derive finite-time bounds of the excess asymptotic variance to zero. We apply our methodology to Stochastic Gradient MCMC (SGMCMC) methods for Bayesian inference on large data sets and combine it with existing variance reduction methods for SGMCMC. We present empirical results carried out on a number of benchmark examples showing that our variance reduction method achieves significant improvement as compared to state-of-the-art methods at the expense of a moderate increase of computational overhead.
Recommendations
- Variance reduction for Markov chains with application to MCMC
- Control variates for stochastic gradient MCMC
- Stochastic gradient Hamiltonian Monte Carlo with variance reduction for Bayesian inference
- Variance reduction for Metropolis-Hastings samplers
- Variance reduction for additive functionals of Markov chains via martingale representations
Cites work
- A new weak dependence condition and applications to moment inequalities
- Analysis and geometry of Markov diffusion operators
- Batch means and spectral variance estimators in Markov chain Monte Carlo
- Bounding \(\bar d\)-distance by informational divergence: A method to prove measure concentration
- Bracketing metric entropy rates and empirical central limit theorems for function classes of Besov- and Sobolev-type
- Consistency and fluctuations for stochastic gradient Langevin dynamics
- Control functionals for Monte Carlo integration
- Control Variates for Estimation Based on Reversible Markov Chain Monte Carlo Samplers
- Control variates for stochastic gradient MCMC
- Exploiting multi-core architectures for reduced-variance estimation with intractable likelihoods
- Exponential convergence of Langevin distributions and their discrete approximations
- Exponential integrability and transportation cost related to logarithmic Sobolev inequalities
- Generalization of an inequality by Talagrand and links with the logarithmic Sobolev inequality
- High-dimensional Bayesian inference via the unadjusted Langevin algorithm
- scientific article; zbMATH DE number 1350773 (Why is no real title available?)
- scientific article; zbMATH DE number 1999206 (Why is no real title available?)
- scientific article; zbMATH DE number 3894218 (Why is no real title available?)
- Informative \(g\)-priors for logistic regression
- Markov Chains
- Monte-Carlo methods and stochastic processes. From linear to non-linear
- On some non asymptotic bounds for the Euler scheme
- Simulation and the Monte Carlo method
- The concentration of measure phenomenon
- Theoretical Guarantees for Approximate Sampling from Smooth and Log-Concave Densities
- Transportation cost for Gaussian and other product measures
- Transportation cost-information inequalities and applications to random dynamical systems and diffusions.
- User-friendly guarantees for the Langevin Monte Carlo with inaccurate gradient
- Variance reduction for Markov chains with application to MCMC
- Variance reduction in Monte Carlo estimators via empirical variance minimization
- Zero variance Markov chain Monte Carlo for Bayesian estimators
Cited in
(7)- Variance reduction for additive functionals of Markov chains via martingale representations
- Empirical variance minimization with applications in variance reduction and optimal control
- Variance reduction for Markov chains with application to MCMC
- Stochastic gradient Hamiltonian Monte Carlo with variance reduction for Bayesian inference
- Control variates for stochastic gradient MCMC
- Stochastic gradient MCMC for state space models
- Diffusion approximations and control variates for MCMC
This page was built for publication: Variance reduction for dependent sequences with applications to stochastic gradient MCMC
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4995114)
