Global Convergence of Stochastic Gradient Hamiltonian Monte Carlo for Nonconvex Stochastic Optimization: Nonasymptotic Performance Bounds and Momentum-Based Acceleration
From MaRDI portal
Publication:5058053
DOI10.1287/opre.2021.2162OpenAlexW3210956727WikidataQ114058136 ScholiaQ114058136MaRDI QIDQ5058053
Xuefeng Gao, Lingjiong Zhu, Mert Gürbüzbalaban
Publication date: 1 December 2022
Published in: Operations Research (Search for Journal in Brave)
Full work available at URL: https://arxiv.org/abs/1809.04618
nonconvex optimizationLangevin dynamicsGibbs samplingempirical risk minimizationstochastic gradient methodsmomentum-based acceleration
Related Items
Swarm gradient dynamics for global optimization: the mean-field limit case, Decentralized Bayesian learning with Metropolis-adjusted Hamiltonian Monte Carlo, Multi-index antithetic stochastic gradient algorithm, Unnamed Item
Cites Work
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Unnamed Item
- Optimization by Simulated Annealing
- Accelerated gradient methods for nonconvex nonlinear and stochastic programming
- Weighted Csiszár-Kullback-Pinsker inequalities and applications to transportation inequalities
- Asymptotics of the spectral gap with applications to the theory of simulated annealing
- Nonstationary Markov chains and convergence of the annealing algorithm
- A strong approximation theorem for stochastic recursive algorithms
- On variance reduction for stochastic smooth convex optimization with multiplicative noise
- Metastability in reversible diffusion processes. II: Precise asymptotics for small eigenvalues
- Isotropic hypoelliptic and trend to equilibrium for the Fokker-Planck equation with a high-degree potential
- The landscape of empirical risk for nonconvex losses
- On sampling from a log-concave density using kinetic Langevin diffusions
- User-friendly guarantees for the Langevin Monte Carlo with inaccurate gradient
- Couplings and quantitative contraction rates for Langevin dynamics
- The geometric foundations of Hamiltonian Monte Carlo
- Behavior of accelerated gradient methods near critical points of nonconvex functions
- Ergodicity for SDEs and approximations: locally Lipschitz vector fields and degenerate noise.
- Choosing Regularization Parameters in Iterative Methods for Ill-Posed Problems
- The computation of averages from equilibrium and nonequilibrium Langevin molecular dynamics
- Complete Dictionary Recovery Over the Sphere I: Overview and the Geometric Picture
- Robust Truncated Hinge Loss Support Vector Machines
- Diffusion for Global Optimization in $\mathbb{R}^n $
- Recursive Stochastic Algorithms for Global Optimization in $\mathbb{R}^d $
- Accelerated Methods for NonConvex Optimization
- An Introduction to Statistical Learning
- Stochastic Processes and Applications
- Convergence rate of Riemannian Hamiltonian Monte Carlo and faster polytope volume computation
- Robust Variable Selection With Exponential Squared Loss
- Ridge Regression: Biased Estimation for Nonorthogonal Problems
- Theoretical Guarantees for Approximate Sampling from Smooth and Log-Concave Densities
- Brownian motion in a field of force and the diffusion model of chemical reactions
- Entropy-SGD: biasing gradient descent into wide valleys
