EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Improving sampling efficacy on high-dimensional distributions with thin high-density regions using Conservative Hamiltonian Monte Carlo

Geoffrey McGregor, Andy T S Wan iD

PNAS Nexus · 30.05.2026 · 5 (6)

DOI: 10.1093/pnasnexus/pgag178 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PNAS Nexus, online erschienen am 30.05.2026. Verfasst von Geoffrey McGregor und 1 weiteren Autor:innen. Beteiligte Einrichtung: Department of Mathematics, University of Toronto, 40 St George Street, Toronto, ON, Canada M5S2E4. Die Arbeit nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: Natural Sciences and Engineering Research Council of Canada, University of Northern British Columbia, University of California, Merced.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Hamiltonian Monte Carlo is a prominent Markov Chain Monte Carlo algorithm, which employs symplectic integrators to sample from high-dimensional target distributions in many applications, such as statistical mechanics, Bayesian statistics, and generative models. However, such distributions tend to have thin high-density regions, posing a significant challenge for symplectic integrators to maintain the small energy errors needed for a high acceptance probability. Instead, we propose a variant called Conservative Hamiltonian Monte Carlo, using R-reversible energy-preserving integrators to retain a high acceptance probability. We show our algorithm can achieve approximate stationarity with an error determined by the Jacobian approximation of the energy-preserving proposal map and its approximation error due to inexact solutions of implicit schemes. Numerical evidence shows improved convergence and robustness over integration parameters on target distributions with thin high-density regions and in high dimensions. Moreover, a version of our algorithm can also be applied to target distributions without gradient information.

Schlagwörter

Stichwörter der Autor:innen

Hamiltonian Monte CarloMarkov Chain Monte Carloapproximate stationarityenergy-preserving integrator

Autor:innen und Institutionen

  • Department of Mathematics, University of Toronto, 40 St George Street, Toronto, ON, Canada M5S2E4.
  • Department of Applied Mathematics, University of California, Merced, 5200 North Lake Road, Merced, CA 95343, USA.

ORCID-Profile: Andy T S Wan

Förderung

  • Natural Sciences and Engineering Research Council of Canada
  • University of Northern British Columbia
  • University of California, Merced

Zitieren

APA

McGregor, G., & Wan, A. T. S. (2026). Improving sampling efficacy on high-dimensional distributions with thin high-density regions using Conservative Hamiltonian Monte Carlo. PNAS Nexus, 5(6). https://doi.org/10.1093/pnasnexus/pgag178

Vancouver

McGregor G, Wan ATS. Improving sampling efficacy on high-dimensional distributions with thin high-density regions using Conservative Hamiltonian Monte Carlo. PNAS Nexus. 2026;5(6). doi:10.1093/pnasnexus/pgag178

BibTeX
@article{McGregor2026,
  author = {McGregor, Geoffrey and Wan, Andy T S},
  title = {Improving sampling efficacy on high-dimensional distributions with thin high-density regions using Conservative Hamiltonian Monte Carlo},
  journal = {PNAS Nexus},
  year = {2026},
  volume = {5},
  number = {6},
  doi = {10.1093/pnasnexus/pgag178}
}
Kostenlos starten →