A Framework for Interdomain and Multioutput Gaussian Processes

TL;DR

This paper introduces a modular, scalable framework in GPflow for interdomain and multioutput Gaussian processes, simplifying implementation and testing for large-scale and deep learning applications.

Contribution

It presents a unified mathematical and software framework for scalable approximate inference in multioutput GPs, integrating interdomain methods and convolutional structures.

Findings

Unified interface for multioutput GP models

Simplifies development of deep GP models

Encourages broader adoption of GPs in large-scale applications

Abstract

One obstacle to the use of Gaussian processes (GPs) in large-scale problems, and as a component in deep learning system, is the need for bespoke derivations and implementations for small variations in the model or inference. In order to improve the utility of GPs we need a modular system that allows rapid implementation and testing, as seen in the neural network community. We present a mathematical and software framework for scalable approximate inference in GPs, which combines interdomain approximations and multiple outputs. Our framework, implemented in GPflow, provides a unified interface for many existing multioutput models, as well as more recent convolutional structures. This simplifies the creation of deep models with GPs, and we hope that this work will encourage more interest in this approach.