Deep convolutional Gaussian processes

TL;DR

This paper introduces deep convolutional Gaussian processes, a Bayesian model combining deep Gaussian processes with convolutional structures, significantly enhancing image classification accuracy on datasets like MNIST and CIFAR-10.

Contribution

It presents a novel deep Gaussian process architecture with convolutional structure, advancing Bayesian image classification methods.

Findings

Over 10% accuracy improvement on CIFAR-10

Significant performance boost over existing Gaussian process models

Effective hierarchical local feature detection

Abstract

We propose deep convolutional Gaussian processes, a deep Gaussian process architecture with convolutional structure. The model is a principled Bayesian framework for detecting hierarchical combinations of local features for image classification. We demonstrate greatly improved image classification performance compared to current Gaussian process approaches on the MNIST and CIFAR-10 datasets. In particular, we improve CIFAR-10 accuracy by over 10 percentage points.