On the Convergence of Stochastic Gradient MCMC Algorithms with High-Order Integrators

Abstract

Recent advances in Bayesian learning with large-scale data have witnessed emergence of stochastic gradient MCMC algorithms (SG-MCMC), such as stochastic gradient Langevin dynamics (SGLD), stochastic gradient Hamiltonian MCMC (SGHMC), and the stochastic gradient thermostat. While finite-time convergence properties of the SGLD with a 1st-order Euler integrator have recently been studied, corresponding theory for general SG-MCMCs has not been explored. In this paper we consider general SG-MCMCs with high-order integrators, and develop theory to analyze finite-time convergence properties and their asymptotic invariant measures. Our theoretical results show faster convergence rates and more accurate invariant measures for SG-MCMCs with higher-order integrators. For example, with the proposed efficient 2nd-order symmetric splitting integrator, the {\em mean square error} (MSE) of the posterior average for the SGHMC achieves an optimal convergence rate of $L^{-4/5}$ at $L$ iterations, compared to $L^{-2/3}$ for the SGHMC and SGLD with 1st-order Euler integrators. Furthermore, convergence results of decreasing-step-size SG-MCMCs are also developed, with the same convergence rates as their fixed-step-size counterparts for a specific decreasing sequence. Experiments on both synthetic and real datasets verify our theory, and show advantages of the proposed method in two large-scale real applications.

NIPS 2015

References in corpus (1)

• The Fundamental Incompatibility of Hamiltonian Monte Carlo and Data Subsampling

Cited by in corpus (18)

• Stochastic Gradient MCMC Methods for Hidden Markov Models
• Particle Optimization in Stochastic Gradient MCMC
• A Convergence Analysis for A Class of Practical Variance-Reduction Stochastic Gradient MCMC
• Stochastic Gradient Monomial Gamma Sampler
• Multi-variance replica exchange stochastic gradient MCMC for inverse and forward Bayesian physics-informed neural network
• Sparse Deep Learning: A New Framework Immune to Local Traps and Miscalibration
• Characterizing Membership Privacy in Stochastic Gradient Langevin Dynamics
• Stochastic Gradient Langevin Dynamics Algorithms with Adaptive Drifts
• Bayesian interpretation of SGD as Ito process
• On Connecting Stochastic Gradient MCMC and Differential Privacy
• Laplacian Smoothing Stochastic Gradient Markov Chain Monte Carlo
• Understanding MCMC Dynamics as Flows on the Wasserstein Space
• Stochastic Gradient MCMC with Stale Gradients
• On Transformations in Stochastic Gradient MCMC
• Differentiable Annealed Importance Sampling and the Perils of Gradient Noise
• Dictionary Learning Strategies for Compressed Fiber Sensing Using a Probabilistic Sparse Model
• Revisiting the Effects of Stochasticity for Hamiltonian Samplers
• Stochastic Approximate Gradient Descent via the Langevin Algorithm