Discontinuous Hamiltonian Monte Carlo for discrete parameters and discontinuous likelihoods

Published

Journal Article

© 2020 Biometrika Trust. Hamiltonian Monte Carlo has emerged as a standard tool for posterior computation. In this article we present an extension that can efficiently explore target distributions with discontinuous densities. Our extension in particular enables efficient sampling from ordinal parameters through the embedding of probability mass functions into continuous spaces. We motivate our approach through a theory of discontinuous Hamiltonian dynamics and develop a corresponding numerical solver. The proposed solver is the first of its kind, with a remarkable ability to exactly preserve the Hamiltonian. We apply our algorithm to challenging posterior inference problems to demonstrate its wide applicability and competitive performance.

Full Text

Duke Authors

Cited Authors

  • Nishimura, A; Dunson, DB; Lu, J

Published Date

  • June 1, 2020

Published In

Volume / Issue

  • 107 / 2

Start / End Page

  • 365 - 380

Electronic International Standard Serial Number (EISSN)

  • 1464-3510

International Standard Serial Number (ISSN)

  • 0006-3444

Digital Object Identifier (DOI)

  • 10.1093/biomet/asz083

Citation Source

  • Scopus