An adaptive Hessian approximated stochastic gradient MCMC method

DOI10.1016/J.JCP.2021.110150MaRDI QIDQ2128489zbMATH OpenOpenAlexFDO

Authors Yating Wang, Wei Deng, Guang Lin

Publication date 22 April 2022

Published in Journal of Computational Physics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2010.01384

zbMATH Keywords

deep learningstochastic approximationlimited memory BFGSadaptive Bayesian methodHessian approximated stochastic gradient MCMChighly correlated density

Mathematics Subject Classification ID

Mathematical programming (90Cxx) Markov processes (60Jxx) Numerical methods for mathematical programming, optimization and variational techniques (65Kxx)

Recommendations

• Bayesian sparse learning with preconditioned stochastic gradient MCMC and its applications
• Neural network gradient Hamiltonian Monte Carlo
• Efficient Bayes inference in neural networks through adaptive importance sampling
• Efficient and generalizable tuning strategies for stochastic gradient MCMC
• scientific article; zbMATH DE number 7307488

Cites work

• A Stochastic Approximation Method
• A mixed multiscale finite element method for elliptic problems with oscillating coefficients
• A stochastic quasi-Newton method for large-scale optimization
• Exploration of the (non-)asymptotic bias and variance of stochastic gradient Langevin dynamics
• Global convergence of online limited memory BFGS
• Langevin diffusions and the Metropolis-adjusted Langevin algorithm
• Mixed generalized multiscale finite element methods and applications
• On the limited memory BFGS method for large scale optimization
• Riemann manifold Langevin and Hamiltonian Monte Carlo methods. With discussion and authors' reply
• SGD-QN: careful quasi-Newton stochastic gradient descent
• Stochastic Quasi-Newton Methods for Nonconvex Stochastic Optimization

Cited in

(3)

• Efficient Bayes inference in neural networks through adaptive importance sampling
• Multi-variance replica exchange SGMCMC for inverse and forward problems via Bayesian PINN
• An adaptive Hessian approximated stochastic gradient MCMC method

Describes a project that uses

Uses Software

• SGD-QN
• L-BFGS

This page was built for publication: An adaptive Hessian approximated stochastic gradient MCMC method

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2128489)