We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cond-mat.stat-mech

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Condensed Matter > Statistical Mechanics

Title: Accelerated Jarzynski Estimator with Deterministic Virtual Trajectories

Abstract: The Jarzynski estimator is a powerful tool that uses nonequilibrium statistical physics to numerically obtain partition functions of probability distributions. The estimator reconstructs partition functions with trajectories of the simulated Langevin dynamics through the Jarzynski equality. However, the original estimator suffers from slow convergence because it depends on rare trajectories of stochastic dynamics. In this paper, we present a method to significantly accelerate the convergence by introducing deterministic virtual trajectories generated in augmented state space under the Hamiltonian dynamics. We theoretically show that our approach achieves second-order acceleration compared to a naive estimator with the Langevin dynamics and zero variance estimation on harmonic potentials. We also present numerical experiments on three multimodal distributions and a practical example where the proposed method outperforms the conventional method, and provide theoretical explanations.
Subjects: Statistical Mechanics (cond-mat.stat-mech); Machine Learning (cs.LG)
Journal reference: Physical Review E, 105, 054120 (2022)
DOI: 10.1103/PhysRevE.105.054120
Cite as: arXiv:2103.00529 [cond-mat.stat-mech]
  (or arXiv:2103.00529v2 [cond-mat.stat-mech] for this version)

Submission history

From: Nobumasa Ishida [view email]
[v1] Sun, 28 Feb 2021 14:59:03 GMT (2681kb,D)
[v2] Wed, 9 Mar 2022 04:26:06 GMT (3107kb,D)

Link back to: arXiv, form interface, contact.