A splitting Hamiltonian Monte Carlo method for efficient sampling

TL;DR

This paper introduces a splitting Hamiltonian Monte Carlo (SHMC) algorithm that improves sampling efficiency for complex systems by splitting potential energy and using random mini-batches, with proven error bounds and verified numerical performance.

Contribution

The paper develops a novel splitting Hamiltonian Monte Carlo method combined with random mini-batch strategies, enabling efficient sampling of systems with singular potentials or multiple barriers.

Findings

Efficient sampling from systems with singular potentials or multiple barriers.

Error bounds of $ ext{O}(rac{1}{ oot{2}\Delta t})$ for the Hamiltonian approximation.

Numerical experiments confirm theoretical error estimates and computational efficiency.

Abstract

We propose a splitting Hamiltonian Monte Carlo (SHMC) algorithm, which can be computationally efficient when combined with the random mini-batch strategy. By splitting the potential energy into numerically nonstiff and stiff parts, one makes a proposal using the nonstiff part of $U$, followed by a Metropolis rejection step using the stiff part that is often easy to compute. The splitting allows efficient sampling from systems with singular potentials (or distributions with degenerate points) and/or with multiple potential barriers. In our SHMC algorithm, the proposal only based on the nonstiff part in the splitting is generated by the Hamiltonian dynamics, which can be potentially more efficient than the overdamped Langevin dynamics. We also use random batch strategies to reduce the computational cost to $\mathcal{O}(1)$ per time step in generating the proposals for problems arising from many-body systems and Bayesian inference, and prove that the errors of the Hamiltonian induced by the random batch approximation is $\mathcal{O}(\sqrt{\Delta t})$ in the strong and $\mathcal{O}(\Delta t)$ in the weak sense, where $\Delta t$ is the time step. Numerical experiments are conducted to verify the theoretical results and the computational efficiency of the proposed algorithms in practice.