A Bayes-Sard Cubature Method
share
This paper focusses on the formulation of numerical integration as an inferential task. To date, research effort has focussed on the development of Bayesian cubature, whose distributional output provides uncertainty quantification for the integral. However, the natural point estimators associated with Bayesian cubature do not, in general, correspond to standard cubature rules (e.g. Gaussian, Smolyak, quasi-Monte Carlo), which are widely-used and better-studied. To address this drawback, we present Bayes-Sard cubature, a general framework in which any cubature rule can be endowed with a meaningful probabilistic output. This is achieved by considering a Gaussian process model for the integrand, whose mean is a parametric regression model with an improper flat prior on each regression coefficient. The features in the regression model consist of test functions which are exactly integrated, with the remainder of the computational budget afforded to the non-parametric part in a manner similar to Sard. It is demonstrated that, through a judicious choice of test functions, any cubature rule can be recovered as the posterior mean in the Bayes-Sard output. The asymptotic convergence of the Bayes-Sard cubature method is established and our theoretical results are numerically verified.
READ FULL TEXT
