Posterior Contraction for Deep Gaussian Process Priors
Gianluca Finocchio, Johannes Schmidt-Hieber; 24(66):1−49, 2023.
Abstract
This study examines the rates of posterior contraction for a class of deep Gaussian process priors in the nonparametric regression setting, given a general composition assumption on the regression function. Our findings demonstrate that the contraction rates can achieve the minimax convergence rate (up to log n factors) and are adaptable to the underlying structure and smoothness of the target function. This proposed framework expands upon the Bayesian nonparametric theory for Gaussian process priors.
[abs]