Sampling algorithms in statistical physics: a guide for statistics and machine learning

Michael F. Faulkner,Samuel Livingstone
2023-06-09
Abstract:We discuss several algorithms for sampling from unnormalized probability distributions in statistical physics, but using the language of statistics and machine learning. We provide a self-contained introduction to some key ideas and concepts of the field, before discussing three well-known problems: phase transitions in the Ising model, the melting transition on a two-dimensional plane and simulation of an all-atom model for liquid water. We review the classical Metropolis, Glauber and molecular dynamics sampling algorithms before discussing several more recent approaches, including cluster algorithms, novel variations of hybrid Monte Carlo and Langevin dynamics and piece-wise deterministic processes such as event chain Monte Carlo. We highlight cross-over with statistics and machine learning throughout and present some results on event chain Monte Carlo and sampling from the Ising model using tools from the statistics literature. We provide a simulation study on the Ising and XY models, with reproducible code freely available online, and following this we discuss several open areas for interaction between the disciplines that have not yet been explored and suggest avenues for doing so.
Computation,Statistical Mechanics
What problem does this paper attempt to address?
The paper primarily explores several algorithms used for sampling from non-standardized probability distributions in statistical physics, and expresses these algorithms in the language of statistics and machine learning. Its core objective is to review and interpret sampling algorithms widely used in statistical physics, while highlighting the connections between these methods and the fields of statistics and machine learning. Specifically, the paper discusses three typical problems: 1. **Phase transitions in the Ising model**: Investigating the phase transition phenomena in 1-dimensional and 2-dimensional Ising models, particularly the transition from paramagnetic to ferromagnetic states as temperature changes under zero external field conditions. 2. **Melting transition on a 2-dimensional plane**: Using the 2-dimensional hard disk model to simulate the phase transition process from solid to liquid, with a particular focus on the impact of density on the phase transition. 3. **Simulation of all-atom water models**: Studying more complex physical interactions through molecular modeling of liquid water. The paper first introduces the basic concepts of statistical physics, including microscopic models, periodic boundary conditions, observables, entropy, free energy, etc., and then focuses on analyzing three representative problems: phase transitions in the Ising model, melting transitions on a 2-dimensional plane, and the construction of all-atom water models. For each problem, the paper provides detailed introductions to the relevant physical background knowledge, the models used, and the potential sampling algorithms. Additionally, it discusses classical sampling algorithms (such as the Metropolis algorithm, Glauber dynamics, and molecular dynamics) as well as some newer methods (such as cluster algorithms, variations of hybrid Monte Carlo, Langevin dynamics, and event-chain Monte Carlo). Finally, the paper offers simulation studies on the Ising model and the XY model, and discusses unexplored interactive areas between statistical physics and machine learning, proposing possible directions for collaboration.