Numerical integration
Bayesian quadrature is a model-based approach to numerical integration, especially efficient for low-dimensional integrals.
Bayesian quadrature is a model-based approach to numerical integration similar to how Bayesian optimization is a model-based approach to global optimization. It is especially efficient for low-dimensional integrals.
Bayesian quadrature is a probabilistic numerical method as it yields a posterior over the integral value, possibly a richer quantity than a point estimate.