Metropolis Monte Carlo sampling: convergence, localization transition and optimality
This work provides insights into optimization and efficiency for researchers using Markov Chain Monte Carlo methods in statistical physics or machine learning, though it appears incremental as it builds on existing models.
The paper studied the convergence properties of Metropolis Monte Carlo sampling, showing that deviations from the target distribution can exhibit a localization transition based on jump length parameters, with relaxation limited by diffusion before the transition and rejection rates after it.
Among random sampling methods, Markov Chain Monte Carlo algorithms are foremost. Using a combination of analytical and numerical approaches, we study their convergence properties towards the steady state, within a random walk Metropolis scheme. Analysing the relaxation properties of some model algorithms sufficiently simple to enable analytic progress, we show that the deviations from the target steady-state distribution can feature a localization transition as a function of the characteristic length of the attempted jumps defining the random walk. While the iteration of the Monte Carlo algorithm converges to equilibrium for all choices of jump parameters, the localization transition changes drastically the asymptotic shape of the difference between the probability distribution reached after a finite number of steps of the algorithm and the target equilibrium distribution. We argue that the relaxation before and after the localisation transition is respectively limited by diffusion and rejection rates.