Deep Structured Mixtures of Gaussian Processes

Abstract

Gaussian Processes (GPs) are powerful non-parametric Bayesian regression models that allow exact posterior inference, but exhibit high computational and memory costs. In order to improve scalability of GPs, approximate posterior inference is frequently employed, where a prominent class of approximation techniques is based on local GP experts. However, local-expert techniques proposed so far are either not well-principled, come with limited approximation guarantees, or lead to intractable models. In this paper, we introduce deep structured mixtures of GP experts, a stochastic process model which i) allows exact posterior inference, ii) has attractive computational and memory costs, and iii) when used as GP approximation, captures predictive uncertainties consistently better than previous expert-based approximations. In a variety of experiments, we show that deep structured mixtures have a low approximation error and often perform competitive or outperform prior work.

arXiv commentAISTATS 2020

References in corpus (7)

• Sum-Product Networks: A New Deep Architecture
• Kernel Interpolation for Scalable Structured Gaussian Processes (KISS-GP)
• Generalized Product of Experts for Automatic and Principled Fusion of Gaussian Process Predictions
• When Gaussian Process Meets Big Data: A Review of Scalable GPs
• Distributed Variational Inference in Sparse Gaussian Process Regression and Latent Variable Models
• Bayesian Learning of Sum-Product Networks
• Optimisation of Overparametrized Sum-Product Networks

Cited by in corpus (2)

• Sequential Gaussian Processes for Online Learning of Nonstationary Functions
• Probabilistic Circuits for Variational Inference in Discrete Graphical Models