Deep Gaussian Processes with Convolutional Kernels

TL;DR

This paper introduces Convolutional Deep Gaussian Processes (CDGPs) that incorporate convolutional kernels to effectively model image data, enabling DGPs to be applied to computer vision tasks with improved performance.

Contribution

The paper develops convolutional kernels for DGPs, allowing them to capture local image features and outperform existing GP baselines in image classification tasks.

Findings

CDGPs outperform GP baselines on image classification

Effective convolutional kernels capture local spatial features

Framework is computationally efficient for image data

Abstract

Deep Gaussian processes (DGPs) provide a Bayesian non-parametric alternative to standard parametric deep learning models. A DGP is formed by stacking multiple GPs resulting in a well-regularized composition of functions. The Bayesian framework that equips the model with attractive properties, such as implicit capacity control and predictive uncertainty, makes it at the same time challenging to combine with a convolutional structure. This has hindered the application of DGPs in computer vision tasks, an area where deep parametric models (i.e. CNNs) have made breakthroughs. Standard kernels used in DGPs such as radial basis functions (RBFs) are insufficient for handling pixel variability in raw images. In this paper, we build on the recent convolutional GP to develop Convolutional DGP (CDGP) models which effectively capture image level features through the use of convolution kernels, therefore opening up the way for applying DGPs to computer vision tasks. Our model learns local spatial influence and outperforms strong GP based baselines on multi-class image classification. We also consider various constructions of convolution kernel over the image patches, analyze the computational trade-offs and provide an efficient framework for convolutional DGP models. The experimental results on image data such as MNIST, rectangles-image, CIFAR10 and Caltech101 demonstrate the effectiveness of the proposed approaches.