Irreversible Monte Carlo Algorithms for Efficient Sampling

Abstract

Equilibrium systems evolve according to Detailed Balance (DB). This principe guided development of the Monte-Carlo sampling techniques, of which Metropolis-Hastings (MH) algorithm is the famous representative. It is also known that DB is sufficient but not necessary. We construct irreversible deformation of a given reversible algorithm capable of dramatic improvement of sampling from known distribution. Our transformation modifies transition rates keeping the structure of transitions intact. To illustrate the general scheme we design an Irreversible version of Metropolis-Hastings (IMH) and test it on example of a spin cluster. Standard MH for the model suffers from the critical slowdown, while IMH is free from critical slowdown.

4 pages, 2 figures

References in corpus (2)

• Event-chain Monte Carlo algorithms for hard-sphere systems
• Distributed Averaging via Lifted Markov Chains

Cited by in corpus (76)

• A Survey of Monte Carlo Methods for Parameter Estimation
• Quantum Speed Limit is Not Quantum
• Generalized event-chain Monte Carlo: Constructing rejection-free global-balance algorithms from infinitesimal steps
• The Mpemba index and anomalous relaxation
• The Zig-Zag Process and Super-Efficient Sampling for Bayesian Analysis of Big Data
• Variance Reduction using Nonreversible Langevin Samplers
• Non-reversible Metropolis-Hastings
• Non-Reversible Parallel Tempering: a Scalable Highly Parallel MCMC Scheme
• Ergodicity of the zigzag process
• Violation of detailed balance accelerates relaxation
• Event-chain Monte Carlo: foundations, applications, and prospects
• A piecewise deterministic scaling limit of Lifted Metropolis-Hastings in the Curie-Weiss model
• Improving the convergence of reversible samplers
• Piecewise Deterministic Markov Processes for Scalable Monte Carlo on Restricted Domains
• Event-chain Monte Carlo for classical continuous spin models
• Event-chain algorithm for the Heisenberg model: Evidence for $z\simeq 1$ dynamic scaling
• Dynamics of the One-Dimensional Ising Model without Detailed Balance Condition
• Speedups in nonequilibrium thermal relaxation: Mpemba and related effects
• Eigenvalue analysis of an irreversible random walk with skew detailed balance conditions
• Sampling efficiency of transverse forces in dense liquids
• Role of current fluctuations in nonreversible samplers
• Efficient Irreversible Monte Carlo samplers
• Thermodynamic bounds on spectral perturbations, with applications to oscillations and relaxation dynamics
• Large deviations for the Skew-Detailed-Balance Lifted-Markov processes to sample the equilibrium distribution of the Curie-Weiss model
• Mathematical understanding of detailed balance condition violation and its application to Langevin dynamics
• Optimization of Markov process violates detailed balance condition
• Geometric Allocation Approaches in Markov Chain Monte Carlo
• Event-chain Monte Carlo with factor fields
• Irreversible simulated tempering
• Mixing and perfect sampling in one-dimensional particle systems
• Accelerated Sampling on Discrete Spaces with Non-Reversible Markov Processes
• Monte Carlo simulations of glass-forming liquids beyond Metropolis
• Irreversible Samplers from Jump and Continuous Markov Processes
• Irreversible Markov chain Monte Carlo algorithm for self-avoiding walk
• Asymmetric Thermal Relaxation in Driven Systems: Rotations go Opposite Ways
• Dynamics of the two-dimensional directed Ising model in the paramagnetic phase
• Lifted Worm Algorithm for the Ising Model
• PDMP characterisation of event-chain Monte Carlo algorithms for particle systems
• Hard-disk dipoles and non-reversible Markov chains
• Anomalous diffusion analysis of the lifting events in the event-chain Monte Carlo for the classical XY models
• Local Equivalence of Reversible and General Markov Kinetics
• Peskun-Tierney ordering for Markov chain and process Monte Carlo: beyond the reversible scenario
• Information Thermodynamics of the Transition-Path Ensemble
• Large-scale dynamics of event-chain Monte Carlo
• Irreversible Boltzmann samplers in dense liquids: weak-coupling approximation and mode-coupling theory
• Accelerating, to some extent, the $p$-spin dynamics
• Solution of the Metropolis dynamics on a complete graph with application to the Markov chain Mpemba effect
• Stochastic gradient method with accelerated stochastic dynamics
• AMAGOLD: Amortized Metropolis Adjustment for Efficient Stochastic Gradient MCMC
• Control of probability flow in Markov chain Monte Carlo -- Nonreversibility and lifting
• Designed-walk replica-exchange method for simulations of complex systems
• Estimating the Mixing Time of Ergodic Markov Chains
• Complexity of zigzag sampling algorithm for strongly log-concave distributions
• Non-reversible Markov chain Monte Carlo for sampling of districting maps
• Direction-sweep Markov chains
• Sampling algorithms in statistical physics: a guide for statistics and machine learning
• Irreversible swap algorithms for soft sphere glasses
• Lifted directed-worm algorithm
• Nonreversible MCMC from conditional invertible transforms: a complete recipe with convergence guarantees
• Rigorous lower bound on dynamical exponents in gapless frustration-free systems
• Deterministic Gibbs Sampling via Ordinary Differential Equations
• Reducing rejection exponentially improves Markov chain Monte Carlo sampling
• Graphical Representations and Worm Algorithms for the O($N$) Spin Model
• Simulated tempering with irreversible Gibbs sampling techniques
• Piecewise deterministic sampling with splitting schemes
• Comparison of the clock, stochastic cutoff, and Tomita Monte Carlo methods in simulating the dipolar triangular lattice at criticality
• Involutive MCMC: a Unifying Framework
• Zigzag path connects two Monte Carlo samplers: Hamiltonian counterpart to a piecewise deterministic Markov process
• Stochastic dynamics without detailed balance condition connecting simple gradient method and Hamiltonian Monte Carlo
• Improving the convergence of Markov chains via permutations and projections
• Quantum Speedups for Multiproposal MCMC
• Steady state distributions of network of degenerate optical parametric oscillators in solving combinatorial optimization problems
• Extended framework for the hybrid Monte Carlo in lattice gauge theory
• Accelerating Multicanonical Sampling with Irreversibility
• Numerical studies for an ab initio investigation into the Boltzmann prescription in statistical mechanics of large systems
• Upper bound inequality for calculation time in simulated annealing analogous to adiabatic theorem in quantum systems