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

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

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

Computer Science > Machine Learning

Title: Revisiting the Effects of Stochasticity for Hamiltonian Samplers

Abstract: We revisit the theoretical properties of Hamiltonian stochastic differential equations (SDES) for Bayesian posterior sampling, and we study the two types of errors that arise from numerical SDE simulation: the discretization error and the error due to noisy gradient estimates in the context of data subsampling. Our main result is a novel analysis for the effect of mini-batches through the lens of differential operator splitting, revising previous literature results. The stochastic component of a Hamiltonian SDE is decoupled from the gradient noise, for which we make no normality assumptions. This leads to the identification of a convergence bottleneck: when considering mini-batches, the best achievable error rate is $\mathcal{O}(\eta^2)$, with $\eta$ being the integrator step size. Our theoretical results are supported by an empirical study on a variety of regression and classification tasks for Bayesian neural networks.
Subjects: Machine Learning (cs.LG); Computation (stat.CO)
Cite as: arXiv:2106.16200 [cs.LG]
  (or arXiv:2106.16200v2 [cs.LG] for this version)

Submission history

From: Giulio Franzese [view email]
[v1] Wed, 30 Jun 2021 16:50:11 GMT (742kb,D)
[v2] Thu, 4 Nov 2021 19:42:54 GMT (1788kb,D)

Link back to: arXiv, form interface, contact.