Metropolis Adjusted Microcanonical Hamiltonian Monte Carlo

Jakob Robnik, Reuben Cohn-Gordon, Uroš Seljak · 2025

Read Metropolis Adjusted Microcanonical Hamiltonian Monte Carlo by Jakob Robnik, Reuben Cohn-Gordon, Uroš Seljak (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-01707.txt.

Description

Jakob Robnik and co-authors present a 2025 sampling method that fixes a bias flaw in a fast Monte Carlo scheme. Sampling from complex distributions is a bottleneck across science. Hamiltonian Monte Carlo is standard but struggles in very high dimensions. A newer microcanonical variant runs faster yet gives biased samples.

Their fix adds a Metropolis-Hastings step to the microcanonical motion. The result, named MAMS, yields unbiased samples in the long run. An auto-tuning scheme sets its knobs with no hand work. Benchmarks show it beats the classic sampler across the board. High-dimensional physics and Bayes problems stand to gain.

Catalog record

CREATORJakob Robnik, Reuben Cohn-Gordon, Uroš Seljak
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE73 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-01707

Browse the full library · About the Tower of Babel