Deep Sigma Point Processes
arXiv:2002.09112
Abstract
We introduce Deep Sigma Point Processes, a class of parametric models inspired by the compositional structure of Deep Gaussian Processes (DGPs). Deep Sigma Point Processes (DSPPs) retain many of the attractive features of (variational) DGPs, including mini-batch training and predictive uncertainty that is controlled by kernel basis functions. Importantly, since DSPPs admit a simple maximum likelihood inference procedure, the resulting predictive distributions are not degraded by any posterior approximations. In an extensive empirical comparison on univariate and multivariate regression tasks we find that the resulting predictive distributions are significantly better calibrated than those obtained with other probabilistic methods for scalable regression, including variational DGPs--often by as much as a nat per datapoint.
arXiv comment15 pages, 13 figures; as appeared in UAI 2020
References in corpus (8)
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- Gaussian Processes for Big Data
- Stochastic Variational Deep Kernel Learning
- Random Feature Expansions for Deep Gaussian Processes
- Inference in Deep Gaussian Processes using Stochastic Gradient Hamiltonian Monte Carlo
- Generalized Variational Inference: Three arguments for deriving new Posteriors
- Orthogonally Decoupled Variational Gaussian Processes
- Robust Deep Gaussian Processes