English

On $L^q$ Convergence of the Hamiltonian Monte Carlo

Classical Analysis and ODEs 2022-06-07 v2 Data Structures and Algorithms Machine Learning Functional Analysis Machine Learning

Abstract

We establish LqL_q convergence for Hamiltonian Monte Carlo algorithms. More specifically, under mild conditions for the associated Hamiltonian motion, we show that the outputs of the algorithms converge (strongly for 2q<2\le q<\infty and weakly for 1<q<21<q<2) to the desired target distribution.

Keywords

Cite

@article{arxiv.2101.08688,
  title  = {On $L^q$ Convergence of the Hamiltonian Monte Carlo},
  author = {Soumyadip Ghosh and Yingdong Lu and Tomasz Nowicki},
  journal= {arXiv preprint arXiv:2101.08688},
  year   = {2022}
}